Printer Friendly

The Solo (Ngandong) Homo erectus assemblage: a taphonomic assessment.


This paper explores the likely taphonomic history of the Ngandong assemblage of Homo erectus. Ngandong is clearly atypical of other discoveries of hominin remains in fluvial deposits because so many crania are present, but so few other skeletal elements. The absence of isolated teeth, mandibles, and the scarcity of post-cranial elements cannot be attributed to lack of searching by the Dutch excavators in 1931-33. Although crania are preferentially transported along stream channels, and often occur as isolated finds in fluvial deposits, it is unlikely that 12 would end up in the same place through the downstream disarticulation of 12 corpses. Predators can also be excluded as responsible for this assemblage. It is concluded that the Ngandong hominins did not share the same taphonomic history as the other mammalian remains, and are thus unlikely to be of the same age. Three explanations are offered. The first, and most parsimonious, is that they were originally buried in a mass drowning event, and then re-exposed; their crania, now disarticulated from their mandibles, were then transported a short distance and reburied. A variant on this scenario is that the Ngandong hominins were deliberately buried near the river, and their skeletons then re-exposed; the crania would then have moved preferentially downstream. Both these scenarios imply that the hominins are older than the rest of the faunal assemblage. A third and probably unlikely possibility is that the hominin remains represent partial burials (or re-burials) in the terrace, and are thus younger. Because it is most unlikely that the hominins remains share the same taphonomic history as the other fossils at Ngandong, it remains undemonstrated that H. erectus persisted in South East Asia after the arrival of H. sapiens.

Keywords: Solo, Ngandong, Homo erectus, fluvial, taphonomy


The fossil-rich terrace deposits near Ngandong were first discovered by Elbert, expedition geologist on the 1908-11 Selenka expedition, and then rediscovered by ter Haar in 1931. The fossil hominins were excavated under the direction of Oppenoorth of the Dutch Geological Survey in 1931-1933 (Oppenoorth 1932, Koenigswald 1933). This excavation dug away a large area, ca. 100 by 50 m (Swisher et al. 1996:1871), of gently sloping to flat land by a bend in the current river. The highly-schematic section by Oppenoorth (1932: 52) indicates that the hominin fossils came from a sandstone layer mixed with marl cobbles up to 2.5 metres from the surface of the profile (see Figure 1), although Koenigswald (1956: 68) describes the hominin fossils as coming from the bottom 2 feet of a layer of sand and gravel. In total, the remains of 12 hominin crania (Solo I-XI; Solo III was later found to represent two individuals; Koenigswald [1951: 215]) as well as two tibiae were recovered; seven of the individuals were regarded as adult, of which two were male, two female, and the rest, indeterminate (Day 1986: 359). There were also >25,000 other mammalian fossils (Koenigswald 1956: 75), most of which have unfortunately been lost. According to Swisher et al. (1996: 1871), another excavation was conducted between 1976 and 1980 by Indonesian researchers, who dug an adjacent area 25 by 14 m and found 1200 fossils, including two further partial hominin crania and some pelvic fragments. In both excavations, most of the fossil vertebrates and all the hominins appear to have been found within 50 cm of the base of a terrace deposit that was up to 3 metres thick and traceable for 15 kin. However, Bartstra et al. (1988) excavated a test pit near the original excavation, and found that fossil bone was irregularly distributed between ca. 1 and 2.5 m from the surface.


The fossil hominin remains from Oppenoorth's excavation were described at length in a monograph by Weidenreich (1951), sadly incomplete at the time of his death, and in another by Santa Luca (1980), who attributed them to a late population of Homo erectus, a conclusion recently reinforced by Anton (2003: 144). According to Koenigswald (1956: 75), the tibiae "did not differ in any way from those of modern man", although Anton (2003: 152), with more data available, concluded that they were typical of the robusticity seen in earlier forms of Homo.

The hominins from Ngandong are regarded as younger than those from Sangiran, Mo(d)jokerto and Trinil, which seems reasonable given their larger cranial capacity of ca. 1090-1231 [cm.sup.3], compared with capacities of 813-1059 [cm.sup.3] for adult individuals from Trinil and the later deposits at Sangiran (see Anton (2002: 311). On account of the associated mammalian fauna, the Ngandong hominins are usually assigned to the Upper Pleistocene. Dating the deposits in which they were found has proved difficult. Uranium-series dates from Ngandong produced age estimates of 31,000 [+ or -] 2000 and 101,000 + 12,000/-10,000 b.p. (Bartstra et al. 1988; Bartstra and Basoeki 1989: 242); these were regarded as tentative, but consistent with an Upper Pleistocene age for the terrace deposits. Recently, Swisher et al. (1996) recovered bovid teeth from a test pit dug near the site of Oppenoorth's excavations and from his layer 2 (the one containing the fossils), and also analysed one bovid tooth from Oppenoorth's excavation of 1932 that was found in the brief period (January 25, 1932 and March 17, 1932; see Koenigswald 1956: 72) between the discovery of Solo IV and V, as well as another bovid tooth from the 1976-80 excavations from which hominins XIV and XV were found. ESR dating of these teeth gave mean ages of 26.9 [+ or -] 2 ka on an early uptake (EU) model, and 46.3 [+ or -] 3.7 ka on a linear uptake (LU) model. On this basis, they suggested that Homo erectus persisted in South East Asia after the arrival of H. sapiens in Australia and Borneo. (The recent discoveries of H. floresiensis on Flores (Brown et al. 2004, Moorwood et al. 2004) imply that there may even have been a third type of hominin in this part of S.E. Asia at this time). Grun and Thorne (1997) challenged the dates of Swisher's team on technical grounds, and also argued that the hominins had been reworked from older deposits: I shall return to this issue later.

The key issue concerning the Ngandong hominins is whether they share the same taphonomic history as the other mammalian fossils in the excavated deposits. If they do not, we need to consider what this might have been, and whether they might be older or younger than the dates suggested by Swisher et al. (1996). We can begin by examining whether they are typical of the type of hominin fossil remains found in fluvial deposits elsewhere.

Taphonomic questions

1. Are the Ngangdong fossil hominin remains typical of other Pleistocene fluvial deposits containing hominin remains?

The main discoveries to date are summarised in Table 1. As a general observation, hominin remains are very rare from Pleistocene fluvial deposits, and only exceptionally is more than one individual represented. Most finds have been of crania or (more rarely) mandibles; post-cranial remains are very rare, although a tibia fragment was found at Sambungmachan, and a clavicle at Narmada in India (Sankhyan 1997). Most of the discoveries listed in Table 1 have also tended to be "one-offs", in the sense that further searching in the same deposits has rarely led to the discovery of other hominin remains. There are a few exceptions: a partial one is Swanscombe, where a third piece of cranium was found in 1955 that amazingly joined the other two, conjoining pieces found at the same locality in 1935 and 1936; even so, all three can be counted as part of just one skeletal element. Saccopastore is another, where a second individual was found in a different layer in the same sequence six years after the first. The Sambungmachan specimens are from a section of a river, and each find should probably be listed separately.

Three of the discoveries listed in Table 1 are clearly anomalous because of the numbers of individuals that are presented. The first is Trinil, which is especially unusual in that five/six femora were found (one in the excavation of 1891/2, and four others in the excavations of 1897 and 1900, but not noticed until 1932 and 1935, when the fossils from Trinil were eventually catalogued (see Shipman 2001: 449-54); the sixth "femur" is probably a piece of elephant bone (Bartsiokas and Day, 1993: 115)). However, the femora are almost certainly not associated with the original skull cap, are probably sapient (Kennedy 1983), and derived from an overlying layer (Bartsiokas and Day 1993, Bartstra et al. 1988, Day 1995). If the hominin evidence from Trinil is restricted to the calotte and two teeth, it then becomes much more like other hominin finds from fluvial contexts. The second is AL-333 at Hadar: this was composed almost entirely of the remains of at least 13 individuals of Australopithecus afarensis, represented by over 200 specimens of most body parts (Johanson et al. 1982: 380). Radosevich et al. (1992: 23) suggest that these individuals may have died during a catastrophic flooding event, after which their skeletons disarticulated and their bones were scattered before being finally buried a few months later.

The third anomalous assemblage is Ngandong, with 14 cranial specimens, two tibiae and a pelvis fragment. An additional peculiarity is that the tibiae are complete. This is odd because first, complete hominin long bones are not found at any other Pleistocene fluvial context (apart those from Trinil that can be discounted); and secondly, Oppenoorth (1932: 52) observed that most bones at Ngandong were broken into fist-sized fragments.

Ngandong is thus highly atypical of other instances where hominin remains have been found in a fluvial context. The nearest point of comparison is AL-333, where roughly the same number of individuals is represented, but with a much wider range of body parts, and almost no other non-hominin fauna. The next question is whether the dominance of crania at Ngandong resulted from inadequate searching for hominin remains by Oppenoorth's team.

2. The dominance of crania: genuine, or the result of inadequate searching?

Swisher et al. (1996: 1871) attribute the lack of post-cranial material to the lack of searching in the original (and now lost) collection. This seems unlikely for three reasons. First, the material from the 1976-80 excavations was presumably examined for hominin remains, but no mandibles, post-cranial remains or isolated teeth were found. Secondly, the excavators did manage to find and recognise post-cranial fragments and isolated teeth of several other mammals, and thus it seems strange that they were myopic when looking for the remains of H. erectus. Thirdly, there is Koenigswald's own testimony. He was present at the excavations in June, September and November of 1933, and was particularly emphatic that the hominin assemblage comprised only skull fragments and two tibiae: "While every part of the skeleton of the various animals was found, particularly of the banteng and the deer, the collection of human remains, which includes only incomplete skulls and tibiae, must be regarded as artificial. We should note here that we searched especially for other parts of the human skeleton and even dispatched a special team to hunt for isolated human teeth, but no trace of such remains was found" (Koenigswald 1951: 216; italics mine). In a later publication, he again stressed that although over 25,000 non-human bones and teeth had been found, "of human remains only a very particular selection whose incidence was certainly not natural ... there were no lower jaws, no teeth, no vertebrae or other bones". In contrast, he noted, some animal skulls were still articulated to part of the vertebral column or retained their mandibles. (Koenigswald 1956: 72-75). It is also surely pertinent that Koenigswald was perfectly capable of identifying primate teeth--only a few years later, he was the first to identify Gigantopithecus blacki in the collections of fossil teeth in Chinese pharmacies in Hong Kong (Koenigswald 1935). The absence of other body parts, contra Swisher et al. (1996), seems genuine, and cannot be attributed to a lack of searching or competence on the part of Koenigswald and his colleagues. The next issue to consider is whether the preponderance of crania resulted from fluvial transport.

3. Fluvial transport of skeletal parts and carcass disarticulation

As the Ngandong hominins were found in a fluvial context, we need to consider how skeletal parts are transported and sorted in river systems. Boaz and Behrensmeyer (1976) demonstrated that the cranium is transported further and faster in a current than the rest of the skeleton. In their flume channel experiments with individual human skeletal items that were placed in a flume flow of 0.31 cm/second, they showed that the cranium, if unbroken, retains a pocket of air that enables it to maintain buoyancy, and it can thus float at roughly the speed of the current and just below the stream's surface. Even if the cranium is fractured and sinks, its roughly spherical shape means that it is easily rolled along a stream's bed. Either way, it has a much higher transport potential than other bones. Coard and Dennell (1995) repeated Boaz and Behrensmeyer's experiments under similar flow conditions by assessing the transport potential of skeletal items of sheep, dog and macaque when they were articulated with their anatomical neighbours (e.g. cranium and mandible(s), or pelvis, femur and tibia) on the grounds that bones usually enter stream channels as partial or complete carcasses, rather than as individual skeletal elements. Whilst the transport potential of elements increased when articulated, heads were still among the most easily transported element. Could the predominance of crania at Ngandgong be explained as the result of their greater capacity for fluvial transport than the rest of the skeleton?

Body-less heads in river channels: A British perspective

In recent years, there has been a lively debate over how 299 crania but only 14 mandibles ended up in the holocene deposits of the River Thames and its tributaries. One suggestion, advocated by Bradley and Gordon (1988), is that because metal objects from the bronze and iron ages have often been found in river deposits, these cranial remains represent votive offerings of individuals that had been previously defleshed, so that the mandible was detached from the cranium. (As an aside about how interpretations of the same type of evidence differ geographically and by generation, there is an interesting narrative contrast here between Bradley and Gordon (1988) and Koenigswald (1956), with the former suggesting that the late prehistoric British made peaceable votive offerings of their dead, whereas the latter portrayed savage "headhunters of Ngandong", who clubbed their victims to death and then devoured their brains). In contrast, Knusel and Carr (1995) pointed out that corpses that end up in rivers will disarticulate, with the head rapidly separating from the rest of the body, often losing its mandible, and eventually being buried a considerable distance downstream as a result of rolling along the channel bed. Crania thus began their downstream journey in an articulated state, and ended it as isolated elements after disarticulation. They thus suggested that "incidental erosion [of burials], accidental drowning or intentional suicide and subsequent fluvial sorting may be a more parsimonious explanation" (ibid. 167). This conclusion is reinforced by a recent elegant study of prehistoric human and ungulate remains that were found during the construction of a dock basin at Preston, Lancashire, in the 1880's (Turner et al. 2002). The human remains comprised 24 crania and one mandible; the ungulate remains were also largely cranial, although some post-cranial material was found. AMS dating of human, aurochs and red deer specimens showed that they dated from the Neolithic to Saxon periods. The predominance of cranial specimens (both human and ungulate) is attributed to the way corpses or carcasses disarticulate once in a fiver channel, as noted above by Knusel and Carr. Emphatically, Turner et al. (ibid. 430) stress "there is nothing inherently odd about recovery of isolated crania without mandibles" (italics theirs'). They also note that body parts can end up 10's of kilometres apart as a result of fluvial action, and without any human involvement. Many of the red deer heads carried an impressive rack of antlers (comparable in size to the complete water-buffalo cranium with its horns 7 feet wide that is illustrated by Koenigswald (1956: 66)). Turner et al. (ibid.) suggested that these came from stags that drowned upstream in their winter territories; as their carcasses drifted downstream, the weight of the antlers would have dragged the head downwards into soft estuarine sediments, and the rest of the carcass would then have separated and floated further downstream. Many of these heads lacked their facial regions, as did the Ngandong water-buffalo specimen.

Fluvial disarticulation is thus a wholly plausible way of explaining the predominance of human, jaw-less crania in British holocene river and estuarine deposits, as well as the bulk of Pleistocene fossil hominin remains in fluvial deposits, even allowing for some bias in recovery towards skulls, particularly from old excavations. It also very neatly explains the isolated occurrences of cranial fragments, the rarity of mandibles, and absence of post-cranial remains from Sangiran. How applicable is this explanation to the Ngandong assemblage?

The most obvious difference is the density of cranial specimens relative to the size of deposit. The -300 crania from the Thames came from several miles of waterway, and at Preston Docks, the 24 crania were found in an area of ca. 1.5 by 1 km (Turner, pets. comm.). In contrast, the 14 crania at Ngandong were found in an area no larger than 50 by 100 metres - <0.3% of the area of Preston Docks. A second difference is that the Thames and Preston Docks crania were found in the downstream parts of the river, whereas the Ngandong ones are from much further upstream. It would seem that either the Solo River at Ngandong acted as an extraordinarily effective trap (for which there is no evidence), or some other explanation is needed.

A second difference is that tibiae appear to have a low transport potential. In the experiments by Boaz and Behrensmeyer (1976), only the proximal tibia was used, and this was identified as a lag component, unlike the acetabulum. Coard and Dennell (1995) found that complete tibiae of sheep, dog and macaque, and complete pelves of dog and macaque, were not transported under similar flow conditions, but those of sheep and dog were moved when articulated with the pelvis and femur. (The lack of movement by the macaque pelvis, femur and tibia may have been caused by the problems of articulating them, and may not be significant; see Coard and Dennell 1995: 446)). It made no difference whether the bones were dry or saturated when they entered the flume channel (Coard 1999). Although complete human tibiae may be more transportable in stream channels than those of sheep, dog and macaque, (especially if still articulated with the rest of the leg), they are clearly far less transportable than crania. What is not known, of course, are the stream conditions at Ngandong when animal remains were being deposited, or the frequency of skeletal elements of the other mammalian species in those deposits: patterns of bone transport and sorting in real fluvial conditions are almost certainly more complex than in laboratory flume experiments.

Because Ngandong is so atypical of other discoveries of hominin remains in fluvial deposits, it is worth considering alternative explanations.

4. Could carnivores have been involved?

Although carnivores are often invoked as explanations for the presence of hominin remains in Pleistocene contexts, it seems unlikely that they were significant at Ngandong. This is because the only large predator that habitually accumulates carcass remains at their feeding and denning locations is the hyaena, and this is not recorded in the Ngangdong fauna. According to Berghe et al. (2001: 388), the only large carnivore in the Ngandong Faunal Stage is the tiger, Panthera tigris. These typically consume their prey at the scene of the kill, sometimes over several days if undisturbed by other competitors and scavengers (Turner and Anton 1997: 182)

An obvious objection to the argument that the tiger was the only large predator at Ngandong is that the Ngandong Faunal Stage (defined largely from the type site) is probably incomplete. Only 13 taxa are recorded, compared with 20 in the earlier Kedung Brubus fauna (0.7 - 0.8 Ma), which samples a similar open woodland environment. The Punung fauna (125 - 60 kya), which samples humid forest conditions, contained 17 taxa (see Berghe et al. 2001: Table 1). As carnivores tend to be under-represented in faunal sequences, the absence of hyaenids in the Ngandong Fauna may simply reflect insufficient fieldwork. During the preceding Kedung Brubus (0.7 - 0.8 Ma) faunal stage, the short-faced hyaena Crocuta brevirostris is present; C. brevirostris is probably the giant hyaenid Pachycrocuta brevirostris, which is recorded in the lower Pleistocene in the Siwailks of Pakistan (Turner 2004) and southern and northern China (Wanpo et al. 1995; Boaz et al. 2000, 2004). As Java lies on the Asiatic side of the Wallace Line, it would have been connected to the South East Asian mainland on numerous occasions during the Pleistocene, and thus open to immigration by a wide range of herbivores and their predators such as hyaenids. Could we therefore argue either they were present in the Ngandong Faunal Stage, or that the Ngandong hominins were reworked, and derived from an earlier hyaenid accumulation?

Either suggestion seems improbable because hyaenid accumulations of bones typically contain large amounts of isolated teeth and post-cranial material, as well as evidence of gnawing. At the Lower Pleistocene fossil locality of Venta Micena, Spain, which was probably accumulated by Pachycrocuta brevirostris, 64.7% of remains were limb bones (of which 34.6% were metapodials), and 12.5% were isolated teeth (Palmqvist et al. 1996). Striations and tooth marks typical of hyaenids were also very frequent. The best example of a hominin assemblage that was formed by hyaenids is from Locality 1, Zhoukoudian, China. Boaz et al. (2000) have recently studied the entire hominin collection of H. erectus from this locality, and concluded that P. brevirostris was the main accumulator and modifier of their bones. The hominin assemblage largely comprised isolated teeth, skull parts and mandibles, with few proximal limb elements, and almost no distal limb elements. 67% of the hominid sample shows bite marks or other carnivore-inflicted damage. The fragmented nature of the hominin assemblage - quite unlike that from Ngandong--is evident from the observation that 51 hominid individuals are represented by 66 separate body part elements (an average of 1.29 per individual) (Boaz et al. 2004).

It is thus highly improbable that the Ngandong hominins were derived from a hyaenid accumulation; first, hyaenid accumulations are not dominated by isolated crania; secondly, hominin remains in carnivore accumulations are usually highly fragmented; and thirdly, there is no evidence of bite marks and gnawing.

A final but remote possibility is that the Ngandong hominins were the victims of crocodiles. These are known to ambush unwary animals in or near rivers, and the Meganthropus III mandible from Sangiran has a puncture wound consistent with a bite from a crocodile or gavial (Koenigswald 1968). Although crocodilians are not reported from Ngandong, gavials (Gavialis bengawanicus) were present at Trinil (Steele 1989: 135). (The large and often lethal salt-water crocodile Crocodylus porosus, often weighing over three tons, would have been absent in the upstream section of the Solo River). Adult crocodilians appear to eat all their prey if undisturbed, and there is no reason why they would leave only crania and a few complete, undamaged leg bones. It also seems improbable that groups of late H. erectus were so oblivious to the dangers of crocodile that 14 could have been taken in one place.

5. Could the Ngandong hominin assemblage indicate cultural practice?

Having so far failed to explain the taphonomic history of the Ngandong hominin assemblage, we might now consider the possibility that the peculiar representation of body parts stems from human practice. We do not need to resurrect Koenigswald's (1956: 75-76) gore-soaked scenario, whereby a group of head-hunters feasted on the brains of their enemies, and then camped on the terrace in the dry season with their trophies. (As he put it: "Perhaps the horde was taken by surprise and fled; perhaps the skulls were put down to mark off the area" (1956:76). The evidence for headhunting was carefully reviewed by Jacob (1972), and found unconvincing: in particular, there was no evidence of cut-marks, any signs of defleshing or human modification of the crania. Instead, we might consider burial practice. Although no burials have yet been associated with Homo erectus, anatomically modern humans (and in western Eurasia, Neanderthals) practised a wide range of methods of disposing of their dead in the last 100-200 kya, and there are no a priori reasons why H. erectus could not have done the same.


Inadequate searching by the excavators for hominin post-cranial material, and the feeding habits of carnivores can be safely excluded from explanations of the Ngandong hominin assemblage. Although there is probably no single explanation of how so many hominin heads and so few other body parts ended up at Ngandong, there are two sets of explanations that might be considered; these, along with that offered by Swisher et al. (1996), are summarised in Figure 2. The first is that the hominins were derived from a source of skeletons further upstream, or a nearby bank, and the second is that they is that they were intrusive, and resulted from the selective burial or reburial of hominin skeletal parts. Each has obvious implications of the dating of the hominin remains, and can be considered in turn.


1. Were the Ngangdong hominins reworked from a source of older burials?

Santa Luca (1980: 9), and Grun and Thorne (1997) suggested that the Ngandong hominins were reworked, and thus older than the terrace deposits in which the rest of the fauna was found. One of Santa Luca's suggestions was that the hominids were buried unsorted, and then later re-exposed and sorted. Neither he nor Grun and Thorne explained adequately why so many crania were redeposited without their mandibles and with so little other post-cranial material. A previous mass drowning event, or a set of deliberate internments would provide such a supply of crania.

On this scenario, the hominin remains could have been derived from a set of burials either further upstream or from a nearby bank that was eroded when the channel migrated laterally. These burials could have been natural (for example, from a previous mass drowning event in which hominins died and were thereafter buried by sediment) or artificial, if we assume that late populations of H. erectus disposed of their dead in the same way as Neanderthals and early populations of anatomically modern humans. Each cranium would then have entered the river already disarticulated from the rest of the body (including the mandible), and been preferentially transported by stream flow before being redeposited at Ngangdong with the rest of the mammalian assemblage.

2. Were the Ngandong hominin remains deliberately buried or reburied in the terrace deposits ?

If the Ngandong hominin remains were deliberately buried (or reburied), there should have been evidence of graves or pits in the excavated terrace deposits. If these were backfilled with the same material, they would not have been easy to see during the excavation. Unfortunately, we have insufficient details about the original excavations to judge the likelihood of such features being detecting or missed. All we know is that the homininds "were found neither in one particular spot nor in a special layer but were irregularly distributed throughout the whole site" (Koenigswald's (1951 : 215). In modern terms, the excavations of 1931-1933 were not of a high standard, and were carried out largely by untrained workmen who had no geological background, and only "the most rudimentary training" in excavation or faunal identification (Santa Luca 1980: 6). It is thus possible that the excavators failed to notice that the hominin remains were intrusive, and in pits. (However, to paraphrase Oscar Wilde, it might be thought an accident to miss one grave pit and a misfortune to miss two, but to miss 14 seems little short of catastrophic.)

Nevertheless, there have been several instances (largely from excavations before 1939) where excavators failed to notice graves that had been dug into Pleistocene deposits. Some examples of where this has occurred are shown in Table 2, and include Olduvai hominin 1 (an iron age burial dug into Bed I), Galley Hill (a bronze age burial dug into a Pleistocene terrace), and Vogelherd Cave (Neolithic and bronze age burials in an upper Palaeolithic layer; see below). The claimed depth of fossil hominin specimens below the modern surface and their apparent association with Pleistocene fossils are no guarantee that they are ancient. In most cases, archaeological sections are drawn at the deepest part of the profile. What we need, and so rarely see, are three-dimensional views of an excavation showing precisely where a discovery was made or a dating sample was taken. The impression is easily given that a specimen or sample came from a layer that was at a uniform depth below the modern surface across the whole of an excavation. In many situations, this is not the case, and the depth below surface can vary considerably, thus raising the possibility that a specimen was found far nearer the surface than implied by the section, and might even have been intrusive.

A good recent example is the Vogelherd Cave near Stetten in Germany, where a partial skeleton was found in the 1930's in an early Upper Palaeolithic early Aurignacian layer ca. 30-36,000 years old. For a long time, this was the only evidence that the early Aurigancian in Central Europe was made by anatomically modern humans. The skeleton was reported as coming from over two metres below the surface in an Upper Palaeolithic layer. Recent AMS dating (Conard et al. 2004) has shown that it is in fact only 3000-5000 years old. Most likely, the human remains originated from intrusive Neolithic burials near the surface of the cave entrance, and some distance from the drawn profile. (Interestingly, the humans remains recorded at the Vogelherd comprised a cranium and mandible (Stetten I), a cranium (Stettin 2), a humerus (Stetten 3), two vertebrae (Stetten 4) and a metacarpus (Stetten 5). If the site had been excavated by untrained workmen (as at Ngangdong), and if the resulting material had not been checked thoroughly, it is easy to see how the vertebrae (burial 4) and metacarpus (burial 5) could have been overlooked. The resulting human assemblage would then have comprised (as at Ngandong) a few crania and an obvious long bone).

An additional point is that morphology and degree of fossilisation are poor indicators of age. As example, the Hahnofersand (Germany) cranial fragment that came from a fluvial gravel deposit was initially dated at 36,000 b.p. and claimed to show a mixture of Neanderthal and anatomically modern human traits. As such, it was quoted as possible evidence of hybridisation between Neanderthals and incoming populations of modern humans. Recent AMS dating has shown that this specimen is probably mesolithic and no more than 7500 years old. The Starosele "Neanderthal" child turned out to be fully modern (Marks et al. 1997). As a salutary reminder of the dangers of assuming that hominin remains from Pleistocene fluvial deposits are in fact Pleistocene, it is worth noting that the entire Late Pleistocene fossil hominin record from Germany has now disappeared into the Holocene as a result of AMS dating (Bronk Ramsey et al. 2002). (The same also happened in Britain, where several early discoveries of human remains in allegedly Pleistocene deposits turned out to be Holocene; see Oakley 1964: 333.)

Overall, there are no a priori reasons why the Ngandong hominins could not have been buried or reburied in the terrace. Their graves might have been missed by the excavators, as on other excavations of Pleistocene deposits in the early 20th century; there is no information on the precise depth below surface of the hominin remains; and morphology is often a poor indicator of age. Even so, the original excavation would have been exceedingly careless to overlook no fewer than 12 grave pits.


Ngandong is no exception to the general rule that fossil assemblages are rarely explicable in terms of one simple set of processes. As Swisher et al. (1997: 1575-6) state, "a mechanism of selectively working and concentrating 15 hominins of one age into a site the size of Ngandong, without reworking older faunal elements, is taphonomically difficult to explain". Given that most of the hominin remains were obtained by excavations conducted over 70 years ago, and from which few details are available, and most of the other non-hominin evidence has been lost, it is unlikely that a definitive answer will emerge without further excavation. Nevertheless, enough has been learnt since Ngandong was first excavated in the 1930's to show that it is anomalous when compared with other discoveries of hominin remains in Pleistocene fluvial context. It is also most unlikely that the predominance of crania, and absence of mandibles, isolated teeth and scarcity of post-cranial resulted from lack of searching by the excavators. Nor are there any grounds for interpreting the Ngandong hominin assemblage as the result of predators, whether tiger, hyaenid, human headhunters, or crocodile. The three simplest explanations are that i) the hominin remains were reworked, and derived either from a previous mass drowning event or ii) a set of deliberate burials; or iii) they were intrusive, and resulted from the selective burial (or reburial) of isolated crania and a few other skeletal elements. Each implies that their age is probably different from the other faunal remains in the channel fill. The most parsimonious explanation is that the hominins were derived from a previous mass drowning event, and the predominance of crania results from their preferential downstream transport. This explanation provides a ready source of crania that could then be concentrated into a site the size of Ngangdong without their mandibles and almost all their post-cranial elements, and without reworking older faunal elements. It also avoids the need to postulate that H. erectus buried its dead (for which there is, as yet, no evidence), and to assume that the excavators were so incompetent that they either could not recognise human remains (apart from crania and tibiae) and/or numerous pits in the areas they were excavating. None of these possibilities, of course, can be resolved without a detailed history of the terrace deposits at Ngandong, and/or direct dating of the hominin remains themselves. Meanwhile, I suggest that the final date of H. erectus in South East Asia remains uncertain, and Swisher et al.'s (1996) conclusion that it persisted after the arrival of H. sapiens in Australia and Borneo requires corroboration from other sources.


It is to be hoped that on-going research at Ngandong by a team directed by Mike Morwood (University of New England) will help clarify the taphonomic and channel history of this important locality.


I am grateful to comments and criticisms from Andrew Chamberlain and Paul Pettitt (Sheffield) and Iain Davidson (University of New England), and to the organisers of the Australian Archaeological Association's meeting at Armidale in December 2004 for allowing me to present an earlier version of this paper. Dr. Robert van de Noort is thanked for assistance over the translation of the Solo sequence by Oppenoorth (1932). I am also grateful to an anonymous reviewer for constructive and helpful criticism. The author also wishes to thank the British Academy for granting him a Research Professorship that relieves him of teaching and administration.


Aldhouse-Green, S., Pettitt, P. and Stringer, C. 1996. Holocene humans at Pontnewydd and Cae Gronw caves. Antiqutiy 70: 444-7.

Anton, S. 2002. Evolutionary significance of cranial variation in Asian Homo erectus. American Journal of Physical Anthropology 118: 301-23.

Anton, S. 2003. Natural history of Homo erectus. Yearbook of Physical Anthropology 46: 126-70.

Ascenzi, A., Biddittu, I., Cassoli, P.F., Segre, A.G. and Segre-Naldini, E. 1996. A calvarium of late Homo erectus from Ceprano, Italy. Journal of Human Evolution 31: 409-23.

Bartsiokas, A. and Day, M.D. 1993. Electron probe energy dispersive X-ray microanalysis (EDXA) in the investigation of fossil bone: the case of Java man. Proceedings of the Royal Society of London, Series B 252:115-23.

Bartstra, G.-J., Soegondho, S. and Wijk, A. van der. 1988. Ngandong man: age and artefacts. Journal of Human Evolution 17: 325-37.

Bartstra, G.-J. and Basoeki. 1989. Recent work on the Pleistocene and the palaeolithic of Java. Current Anthropology 30:241-4.

Berghe, G.D. van den, Vos, J. de and Sondaar, P.Y. 2001. The Late Quaternary palaeogeography of mammal evolution in the Indonesian Archipelago. Palaeogeography, Palaeoclimatology, Palaeoecology 171: 385-408.

Boaz, N.T. and Behrensmeyer, A.K. 1976. Hominid taphonomy: transport of human skeletal parts in an artificial fluviatile environment. American Journal of Physical Anthropology 45: 53-60.

Boaz, N.T., Ciochon, R.L., Xu Qinqi and Jinyi Liu. 2000. Large mammalian carnivores as a taphonomic factor in the bone accumulation at Zhoukoudian. Acta Anthropologica Sinica 19 (supplement): 224-34.

Boaz, N.T., Ciochon, R.L., Qinqi Xu and Jinyi Liu 2004. Mapping and taphonomic analysis of the Homo erectus loci at Locality I Zhoukoudian, China. Journal of Human Evolution 46: 519-49.

Bradley, R. and Gordon, K. 1988. Human skulls from the river Thames, their dating and significance. Antiquity 62: 503-9.

Bronk Ramsey, C., Higham, T.F.G., Owen, D.C., Pike, A.W.G. and Hedges, R.E.M. 2002. Radiocarbon dates from the Oxford AMS system: Archaeometry datelist 31. Archaeometry 44 (3) Supplement 1: 1-149.

Brown, P., Sutkina, T., Morwood, M.J., Soejono, R.P., Jatniko and Saptomo, E.W. 2004. A new small-bodied hominin from the Late Pleistocene of Flores, Indonesia. Nature 431: 1055-68.

Coard, R. 1999. One bone, two bones, wet bones, dry bones: transport potentials under experimental conditions. Journal of Archaeological Science 26: 1369-75.

Coard, R. and Dennell, R.W. 1995. Taphonomy of some articulated skeletal remains: transport potential in an artificial environment. Journal of Archaeological Science 22: 441-8.

Conard, N.J., Grootes, P.M. and Smith, F.H. 2004. Unexpectedly recent dates for human remains from the Vogelherd. Nature 430: 198-200.

Day, M.H. 1986 Guide to Fossil Man (4th edition). London: Cassell.

Day, M. 1995. Continuity and discontinuity in the postcranial remains of Homo erectus. In H. Ulrich (ed.), Man and the environment in the Palaeolithic. ERAUL 62: 181-90. Liege.

Grun, R. and Thorne, A. 1997. Dating the Ngandong humans. Science 276: 1575.

Jacob, T. 1972. The problem of head-hunting and brain-eating among Pleistocene men in Indonesia. Archaeology and Physical Anthropology in Oceania 7:81-91.

Johanson, D.C., Taieb, M. and Coppens, Y. 1982. Pliocene hominids from the Hadar Formation, Ethiopia (1973-1977): stratigraphic, chronologic and palaeoenvironmental contexts, with notes on hominid morphology and systematics. American Journal of Physical Anthropology 57: 373-402.

Kennedy, G.E. 1983. Some aspects of femoral morphology in Homo erectus. Journal of Human Evolution 12: 587-616.

Knusel, C.J. and Carr, G.C. 1995. On the significance of the crania from the River Thames and its tributaries. Antiquity 69:162-9.

Koenigswald, R. von 1933. Ein neuer Urmenscb aus dem Dilivium Javas. Zbl. Miner Geol., A. und B Palaont. 29-42.

Koenigswald, G.H.R. von. 1935. Eine fossile Saugertierefauna mit Simia aus Sudchine. Proceedings Koninklijke Nederlandse Akademie Wetenschappen Amsterdam 38: 872-9.

Koenigswald, G.H.R. von. 1951. Introduction. Anthropological Papers of the American Museum of Natural History, New York 43: 211-21.

Koenigswald, G.H.R. von. 1956. Meeting Prehistoric Man. London: Thames and Hudson.

Koenigswald, G.H.R. von. 1968. Observations upon two Pithecanthropus mandibles from Sangiran, Central Java. Proceedings of the Academy of Science, Amsterdam B71: 99-107.

Marks, A.E., Demidenko, Y.E., Monigal, K., Usik, V.I., Ferring, C.R., Burke, A., Rink, J. and McKinney, C. 1997. Starosele and the Starosele Child: new excavations, new results. Current Anthropology 38: 112-23.

Morwood, M.J., Soejono, R.P., Roberts, R.G., Sutnika, T., Turney, C.S.M., Westaway, K.E., Rink, W.J., Zhao, J.-X., Bergh, G.D. van den, Due, R.A., Hobbs, D.R., Moore, M.W., Bird, M.I. and Fifield, L.K. 2004. Archaeology and age of a new hominin from Flores in eastern Indonesia. Nature 431: 1087-91.

Oakley, K.P. 1964. Frameworks for Dating Fossil Man. London: Weidenfeld and Nicolson.

Oakley, K., Campbell, B.G. and Molleson, T.V. 1971. Catalogue of Fossil Hominids Part II: Europe. London: British Museum (Natural History).

Oakley, K., Campbell, B.G. and Molleson, T.V. 1975. Catalogue of Fossil Hominids Part III: Americas, Asia, Australasia. London: British Museum (Natural History).

Oppenoorth, W.EE 1932. Homo (Javanthropus) soloensis: een pleistocene Mensch van Java. Scientific Proceedings of the Mining Company of the Dutch East Indies 20: 49-75.

Palmqvist, P., Martinez-Navarro, B. and Arribas, A. 1996. Prey selection by terrestrial carnivores in a lower Pleistocene paleocommunity. Paleobiology 22: 514-34.

Radesovich, S.C., Retallack, G.J. and Taieb, M. 1992. Reassessment of the palaeoenvironment and preservation of hominid fossils from Hadar, Ethiopia. American Journal of Physical Anthropology 87: 15-27.

Sankhyan, A.R. 1997. Fossil clavicle of a Middle Pleistocene hominid from the Central Narmada Valley, India. Journal of Human Evolution 32: 3-16.

Santa Luca, A.P. 1980. The Ngandong fossil hominids: a comparative study of a Far Eastern Homo erectus group. Yale University Publications in Anthropology 78.

Shipman, P. 2001. The Man Who Found the Missing Link: The Extraordinary Life of Eugene Dubois. London: Weidenfeld and Nicholson.

Sonkalia, A. 1985. Skull cap of an early man from the Narmada Valley alluvium (Pleistocene) of Central India. American Anthropologist 87: 612-6.

Steele, R. 1989. Crocodiles. London: Christopher Helm.

Swisher, G.C., Rink, W.J., Anton, S.C., Schwarcz, H.P., Curtis, G.H., Suprijo, A. and Widiasmoro 1996. Latest Homo erectus of Java: potential contemporaneity with Homo sapiens in Southeast Asia. Science 274: 1870-4.

Swisher, G.C., Rink, W.J., Schwarcz, H.P., Anton, S.C. 1997. Response to Dating the Ngandong humans by Grun, R. and Thorne, A. Science 276: 1575-6.

Turner, A. 2004. Carnivore remains from the Pabbi Hills. In R. W. Dennell, Early hominin landscapes in the Pabbi Hills, northern Pakistan. British Archaeological Reports, International Series 1265: 404-11.

Turner, A. and Anton, M. 1997. The Big Cats and their Fossil Relatives. New York: Columbia University Press.

Turner, A., Gonzalez, S. and Ohman. J.C. 2002. Prehistoric human and ungulate remains from Preston Docks, Lancashire, U.K: problems of river finds. Journal of Archaeological Science 29: 423-33.

Wanpo, H., Ciochon, R., Yumin, G., Larick, R., Qiren, F., Schwarcz, H., Yonge, C., de Vos, J. and Rink, W. 1995. Early Homo and associated artefacts from Asia. Nature 378: 275-8.

Weidenreich, F. 1951. Morphology of Solo Man. Anthropological Papers of the American Museum of Natural History, New York 43: 222-90.

Dept. of Archaeology, University of Sheffield, Northgate House, West Street, Sheffield S1 4ET, U.K.
Table 1. Hominin skeletal remains from fluvial deposits.

Site                         Date              Element(s)

Ngandong (Solo), Indonesia   1931-3, 1976-80   14 crania, 2 tibiae,
                                                 1 innominate fragment
Trinil, Indonesia            1891-1900         Calotte, 2 teeth and
                                                 perhaps 5 femora
Mauer, Germany               1907              Mandible
Saccopastore I               1929              Cranium
Saccopastore II              1935              Cranium
Steinheim, Germany           1933              Calvaria
Swanscombe, U.K.             1935-6, 1955      3 conjoining skull
Olduvai OH9, Tanzania        1960              Calvarium
Lantian, China               1963              Mandible
Omo SL7A, Ethiopia           1967              Mandible
Sambungmachan, Indonesia     1973 onwards      3 calvaria, 1 tibia
Hahnofersand, Germany        1973              Skull fragment
Binshof, Germany             1974              Skull
Hadar, AL-333                1975-1977         13 individuals; >200
Paderborn, Germany           1976              Skull fragment
Dali, China                  1978              Cranium
Awash, Ethiopia              1981              Femur and frontal
Narmada, India               1982              Skull fragment,
Ceprano, Italy               1994              Skull fragment

Site                         Age

Ngandong (Solo), Indonesia   Late Pleistocene
Trinil, Indonesia            Early Pleistocene
Mauer, Germany               Middle Pleistocene
Saccopastore I               Upper Pleistocene, level 5
Saccopastore II              Upper Pleistocene, level 7
Steinheim, Germany           Middle Pleistocene
Swanscombe, U.K.             Middle Pleistocene
Olduvai OH9, Tanzania        Lower Pleistocene
Lantian, China               Lower/Middle Pleistocene
Omo SL7A, Ethiopia           Late Pliocene
Sambungmachan, Indonesia     Middle or Upper Pleistocene
Hahnofersand, Germany        36,300 [+ or -] 600 BP
Binshof, Germany             21,300 [+ or -] 320 BP
Hadar, AL-333                Pliocene
Paderborn, Germany           27,400 [+ or -] 600 BP
Dali, China                  Middle Pleistocene
Awash, Ethiopia              Late Pliocene
Narmada, India               Middle Pleistocene
Ceprano, Italy               Middle Pleistocene

Sources: Day, 1986, except for: Sankhyan 1997 and Sonkalia 1985
for Narmada; Johanson et al. 1982 for AL-333; Bronk Ramsey et
al. 2002 for Hahnofersand, Paderborn and Binshof; these are
included because they were thought to be Pleistocene in age when
discovered. Ascenzi et al. (1996) for Ceprano; Oakley et al. (1971:
254) for Saccopastore; and Oakley et al. (1975: 79) for Lantian.
Notes: The Awash specimens were surface finds; the context is
problematic at Sambungmachan. The Hoino ergaster skeleton
WT15000 (Kenya) is excluded as that was derived from the edge
of a swamp, not a river. Two notorious fakes from fluvial contexts
have been excluded but are otherwise consistent with the above:
the Moulin Quignon mandible of 1863, and the Piltdown skull cap
and mandible of 1913-15.

Table 2. Hominin skeletal remains from that have been redated
from the Pleistocene to Holocene.

Site                     Date      Deposit   Specimen

Bury St. Edmonds, U.K.   1882      Fluvial   Calotte
Tilbury, U.K.            1883      Fluvial   Skeleton
Galley Hill, U.K.        1888      Fluvial   Skeleton
Baker's Hole, U..K.      1902      Fluvial   Calvarium
Lloyd's, London          1925      Fluvial   Calotte
Pontnewydd, U.K.         various   Cave      Teeth
Vogelherd, Germany       1931      Cave      Crania, some post-cranial
Hahnofersand, Germany    1973      Fluvial   Skull fragment
Binshof, Germany         1974      Fluvial   Skull
Paderborn, Germany       1976      Fluvial   Skull fragment
Staroselye, Crimea       1953      Cave      Child's skeleton
Olduvai, Tanzania        1913      Bed II    Skeleton

Site                     Initial age estimate   Current age estimate

Bury St. Edmonds, U.K.   Palaeolithic           Bronze age
Tilbury, U.K.            Palaeolithic           Holocene
Galley Hill, U.K.        Middle Pleistocene     3310 [+ or -] 150 BP
Baker's Hole, U..K.      Middle Palaeolithic    Holocene
Lloyd's, London          Palaeolithic           Holocene
Pontnewydd, U.K.         Neanderthal            Holocene
Vogelherd, Germany       30-36,000 BP           3900-5000 BP
Hahnofersand, Germany    36,300 + 600 BP        7500 [+ or -] 55 BP
Binshof, Germany         21,300 + 320 BP        3090 [+ or -] 45 BP
Paderborn, Germany       27,400 + 600 BP        238 [+ or -] 39 BP
Staroselye, Crimea       >40,000 BP             18th century AD
Olduvai, Tanzania        Lower Pleistocene      Iron Age

Sources: As for Table 1, and also Marks et al. 1997 for Staroselye,
Aldhouse-Green et al. 1996 for Pontnewydd, and Oakley, 1964: Table
16 for Galley Hill. Oakley (1964, Table 16) also lists several
other human remains found in the 19th and early 20th century that
have been redated from the Pleistocene to the Holocene.
COPYRIGHT 2005 Blackwell Publishing Limited, a company of John Wiley & Sons, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2005 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Dennell, R.W.
Publication:Archaeology in Oceania
Geographic Code:80OCE
Date:Oct 1, 2005
Previous Article:Tortes Strait archaeology and material culture.
Next Article:A fourteenth-century house from the Rakaia River Mouth, Canterbury, New Zealand.

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |