The Niah Caves

The Niah Caves are a series of massive interconnected caverns and smaller caves located in a limestone massif about 15 km from the north Borneo coast (E. 113° 45′ 160.9′′/N. 3° 48′ 10.19′′). The roughly circular massif is 4 km in diameter and its highest peak (Gunung Subis) is 396 m above sea level. Weathering and erosion have created a dramatic karst scenery of vertical cliffs rising up to narrow plateaus which are separated by deep gorges (Wall Reference Wall1967). The surrounding area comprises undulating marl deposits. The contrast between the high, well-drained, limestone massif and the surrounding low-lying and poorly-drained swampland is reflected in the vegetation of the area (Hazenbroek & Morshidi Reference Hazenbroek and Abang Morshidi2000). The cliff edges and peaks of Gunung Subis have arid tolerant plants and succulents. The surrounding low-lying land is covered in dipterocarp rainforest along its higher drier ridges, with swamp-forest covering the areas in between.

The largest cave system in the massif is an interconnected system of cave passages, chambers, and cave mouths, within a block of limestone, 100 m out from the north-eastern corner of the larger massif. This gigantic cave system is called the Niah Cave Complex; the largest opening of which is the West Mouth. The West Mouth is approximately 160.4 ft (c. 50 m) above sea level (Jee Reference Jee1959) and measures 160 m wide and approximately 60 m high (Fig. 2). The cave entrance is formed by the convergence of two large cave passages, northern and southern. The northern area, where the Neolithic cemetery is located, is formed of a rising pile of guano which provides easy access into the inner system of cave chambers and passages which emerge at five other cave mouths.

Fig. 2

The West Mouth of Niah Cave, looking north-west. The Neolithic cemetery is located in the top right of the picture, adjacent to the cave wall and behind the fencing protecting the archaeological site. (photograph: G. Barker)

Neolithic archaeology and environment

Excavations at numerous caves sites in the 1950s and '60s revealed a long sequence of Pleistocene and Holocene occupation and burial dated from around 40,000 years ago to the present (B. Harrisson Reference Harrisson1967; T. Harrisson Reference Harrisson1957; Reference Harrisson1958; Reference Harrisson1959). The discoveries in the most intensively investigated entrance, the West Mouth, included 25 human burials dating to the Early Holocene, and over 160 Neolithic burials, comprising respectively the largest Mesolithic and Neolithic cemeteries in Island South-east Asia (Figs 3, 4). The programme of renewed fieldwork by the NCP, augmented by an extensive re-dating and archival study, has been able to successfully reconstruct the stratigraphy and occupation history of the site (Barker Reference Barker2013).

Fig. 3

Plan of the archaeological site in the West Mouth of Niah Cave showing locations of the later prehistoric burials, the main concentration of the ceramic assemblage, and burials discussed in the introductory sections of the paper (illustration: L. Lloyd-Smith)

Fig. 4

Excavation in the Neolithic cemetery in 1965. The cemetery started behind the slanting walkway in the middle of the picture. The positions of burials were marked by white flags (photograph courtesy of the Sarawak Museum)

Taking the presence of charcoal as a proxy for human activity in the cave, the new dates suggests that the Early Holocene use of the West Mouth (c. 10,000–6000 bc) was followed by a gap of some 4000 years (Krigbaum Reference Krigbaum2001; Lloyd-Smith Reference Lloyd-Smith2009, Reference Lloyd-Smith2013). The hiatus in cave use between 6000 and 2000 bc coincides with the mid-Holocene high sea stand identified in two pollen cores taken near the caves which have dark bluish–grey laminated clays at the bottom indicative of tidal mangrove swamps (Hunt & Rushworth Reference Hunt and Rushworth2005). At a transition dated to 5710 ± 80 bp or 4721–4367 cal. bc (Beta-193909) in one core and to 5160 ± 60 bp or 4225–3793 cal. bc (Beta-193910) in the other, these pass into dark brown peaty clays and then into dark brown wood peat, indicative of less saline environments characterised by back mangrove swamp. The more open vegetation that then developed around the caves was associated with repeated evidence for localised burning suggestive of human clearance activities (Hunt & Rushworth Reference Hunt and Rushworth2005).

From inside the cave, fragments of domesticated rice have been found as incidental inclusions in earthenware, dated by optically stimulated luminescence to between 1000 bc and ad 200 (Doherty et al. Reference Doherty, Beavitt and Kurui2000). Other archaeobotanical and archaeozoological data from Niah Cave securely associated with the Neolithic phase are sparse. Deposits excavated by the NCP did not contain the remains of any domesticated plant species (Reference Barton, Victor and CarlosBarton et al. forthcoming). This reinforces the impression that during the Neolithic the cave was not used for habitation or food preparation, but was solely reserved as a place for the burial of the dead.

Artefacts obtained from the cemetery area included a substantial quantity of earthenware (over 18,000 sherds, weighing nearly 200 kg; Reference Cole and DalyCole & Daly forthcoming), and an unknown number of whole and fragmented polished stone adzes. Very little of the ceramic assemblage was found in direct association with burials. Only one extended burial is definitely associated with a complete vessel. Twelve burials were contained in earthenware vessels (Lloyd-Smith & Cole Reference Lloyd-Smith and Cole2010). The distribution of sherds appears to be concentrated in an area with very few burials, just to the north-west, towards the cave mouth, of the cemetery. Edge abrasion of the sherds is minimal and there are few use-wear signatures, such as burning or residues, indicative of domestic usage indicating that vast majority of the ceramic assemblage is related to funerary practices (Cole Reference Cole2007). Small vessel forms include open bowls, globular ‘cooking-pots’, bottles, and carinated pots. Medium and large vessel forms include ‘cooking’ and ‘storage’-type vessels, funnel-necked vessels with globular bases, and the unusual double-spouted and three-colour-ware vessels for which Niah has become renowned (Reference Cole and DalyCole & Daly forthcoming; T. Harrisson Reference Harrisson1971; Solheim et al. Reference Solheim, Harrisson and Wall1959).

One noteworthy artefact discovered in the 1977 excavations is a ling-ling-o jade pendant found ‘close to a group of [ceramic] sherds about 7” above a body in grid MD3’ (Fig. 5; Zuraina Majid Reference Zuraina1982, 46). This probably overlay burial B226. Mineralogical analysis identified the jade as comparable to that from a major deposit at Fengtian on eastern Taiwan, c. 2500 km from Niah (Iizuka et al. Reference Iizuka, Bellwood, Datan and Hung2005). Similar pendants have been found in Late Neolithic and Early Metal Age contexts at Tabon Caves on Palawan (Fox Reference Fox1970), Arku Cave on Luzon (Thiel Reference Thiel1990), and many Sa-Huynh sites in southern Vietnam where they are dated to 500–0 bc (Iizuka et al. Reference Iizuka, Bellwood, Datan and Hung2005, 27). The Niah Cave example is the only one yet found on Borneo.

Fig. 5

Jade ling-ling-o pendant found c. 6 in (15 cm) above extended burial B226 (photograph courtesy of the Sarawak Museum)

Physical anthropological data

Morphometric analysis of 15 Early Holocene and 28 Neolithic burials indicates that, although separated by several thousand years, the two populations were fundamentally of the same physical type, with no statistical differences in the upper and mid-face datasets (Manser Reference Manser2005). A significant reduction in the size of the teeth in the Neolithic skeletons compared with those of the Early Holocene was interpreted in terms of dietary change, possibly associated with changes in food preparation such as new methods of cooking foods in ceramic vessels.

In terms of inferred diet, light isotopic analysis indicates broad changes between the Early Holocene and Neolithic populations (Krigbaum Reference Krigbaum2001; Reference Krigbaum and Manser2005). A shift to less negative δ¹³C values in Neolithic burials is taken as indicating that the majority of the diet came from horticultural/agricultural practices in open environments (Krigbaum Reference Krigbaum2005, 83–4). This correlates well with the pollen and sedimentary data from outside the cave (Hunt & Rushworth Reference Hunt and Rushworth2005). At the same time, the Neolithic cemetery population as a whole exhibits a wide range of δ¹³C values and a degree of heterogeneity which suggests a range of dietary regimes, including hunting and gathering, were important (Krigbaum Reference Krigbaum2005, 84). Krigbaum's study employed the original classification of burial types and although reclassification (see below) does not alter the overall conclusions, the new detailed chronology of the Neolithic cemetery illuminates important temporal and spatial patterning in inferred dietary regimes and possible correlations to mortuary practice which are discussed below.

An analysis of strontium and lead isotopes from 34 burials in the Neolithic cemetery identified clusters of isotopic ratios interpreted as different dietary catchments (Valentine et al. Reference Valentine, Kamenov and Krigbaum2008). The strontium (⁸⁷Sr/⁸⁶Sr) distribution of the Niah samples is ‘heavily skewed with three peaks’ (Valentine et al. Reference Valentine, Kamenov and Krigbaum2008, 1469). The first and largest peak is comparable to the value of modern sea-water, and when plotted against the lead (²⁰⁶Pb/²⁰⁴Pb) ratios, sub-divides into two distinct, statistically significant, clusters which may represent two settlement areas with coastal/marine influenced dietary regimes. The two lower value strontium peaks may represent more inland dietary catchment areas.

Three of the West Mouth burials (B3, B36, and B160) had significantly different heavy isotopic ratios suggesting non-local origins for these individuals. The signature from an adult male in multiple secondary burial B160 was comparable to samples analysed from Metal Age burials at the site of Lobang Angin, an inland cave site located in the Mulu Caves c. 120 km to the north-east from Niah (Fig. 1). Similarities in distinctive double-spouted vessels and the three-colour-wares, found both at Niah and Lobang Angin (Datan Reference Datan1993; T. Harrisson Reference Harrisson1971; Solheim et al. Reference Solheim, Harrisson and Wall1959), support the possibility that the adult individual in B160 may have originated from the Mulu area (Valentine et al. Reference Valentine, Kamenov and Krigbaum2008).

Mortuary Practice

By the end of the 1967 season at Niah, 181 features allocated burial numbers had been recorded and fully excavated. The intention had been to return for one final season and complete the excavation of the cemetery, however, this never took place and 33 burials (all in the Neolithic cemetery) were left exposed in situ at the site. In the original classification seven burial types were defined by their ‘main physical outline’ (B. Harrisson Reference Harrisson1967, 132). These were: flexed, seated, mutilations, extended, multiple, and secondary burnt and cremated burials (as well as ‘disturbed and doubtful fragments’, ‘timber only’, and one ‘shell and stone offering’). Burial types were grouped into the Mesolithic (flexed, seated, and mutilations) or the Neolithic (extended, multiple, burnt, and cremated burials) depending on their location and depth. However, anomalies to such clean cut phasing were noted (B. Harrisson Reference Harrisson1967).

In 1977, 11 Neolithic burials were exposed (Zuraina Majid Reference Zuraina1982, 46), one of which, now allocated number B232, was lifted in 2004 (Zuraina Majid et al. Reference Zuraina, Lim, Arif, Samsuddin, Nizam, Zafarina and Norazmi2005). Most recently, 19 burials were investigated by the NCP, including seven already exposed; eight burials were completely lifted, the others were only partially exposed and were reburied (Barker et al. Reference Barker, Badang, Barton, Beavitt, Bird, Daly, Doherty, Gilbertson, Glover, Hunt, Manser, McLaren, Paz, Pyatt, Reynolds, Rose, Rushworth and Stephens2001; Reference Barker, Barton, Bird, Cole, Daly, Gilbertson, Hunt, Krigbaum, Lampert, Lewis, Lloyd-Smith, Manser, Menotti, Paz, Piper, Pyatt, Rabett, Reynolds, Stephens, Thompson, Trickett and Whittaker2002b; Reference Barker, Barton, Bird, Cole, Daly, Dykes, Gilbertson, Hunt, Lewis, Lloyd-Smith, Manser, McLaren, Menotti, Paz, Piper, Pyatt, Rabett, Reynolds, Stephens, Thompson and Trickett2003).

To investigate the spatial and stratigraphic anomalies in the original burial types, a refined classification hierarchically groups all burials from all periods by a number of variables (Table 1; Lloyd-Smith Reference Lloyd-Smith2011). The term ‘burial’ is taken as the deliberate discrete deposition of human remains. In the combined number series there are currently 262 allocated ‘burial’ numbers. After reassessing the burial data, documentary records and, where possible, the skeletal remains, 209 ‘proper’ burials were identified: 25 date to the Early Holocene (Lloyd-Smith Reference Lloyd-Smith2013) and 184 to the Neolithic or later. The main difference between the original and refined classifications is the dispensation of the term ‘mutilations’, which are now defined as types of secondary burial. Many of the ‘disturbed and doubtful fragments’ were also reclassified as proper burials.

Table 1

Classification of burial types in the west mouth, niah cave

The numbers in square brackets denote Early Holocene burials; total 25. Curved brackets denote burials that date to the Neolithic (170) or Post-Neolithic (14); total 184. This includes all possible burials even when remains have not been relocated. The numbers after ‘Arm positioning, ‘Skull only’, and ‘Charred’ in the third order classification column denote sub-sets of burial types not independent numbers of burials

The first order distinction is between primary and secondary burial. Primary burial refers to a single stage funeral for the disposal of the dead, in this case by interment in the ground. Secondary burial implies the transformation of the body, for example through cremation or the exhumation and manipulation of the skeletal remains, followed by a final (secondary) burial. Second order variables include the positioning of primary burials (flexed or extended; Fig. 6) and whether secondary remains were unburnt, cremated, or a mix of semi-articulated and disarticulated remains (Fig. 7). Two oversized coffins, containing primary and secondary burials, retain their original designation as ‘multiple’ burials, as this is felt a useful defining character of these two unique burials (Fig. 8). Third order variables describe further physical manipulation of the burials, in particular the arrangement of the limbs of primary flexed and extended burials, and the selection of skeletal elements (skull) in some secondary burials.

Fig. 6

Primary extended burial B133 with the right arm folded up and left hand over the pelvis (Arm Position D). A naturally shaped oval stone pounder, described by the original excavator as a ‘phallic’ pebble, lies above the head, and two shell or bone ear-rings lie next to left side of the chin (photograph courtesy of the Sarawak Museum)

Fig. 7

Semi-articulated unburnt secondary burial B153. Above: the upper layer of disarticulated remains; below: the underlying articulated segments of the spinal column (photographs courtesy of the Sarawak Museum)

Fig. 8

Log coffin with protruding handle containing multiple burial B138, comprising both primary and secondary burial remains (photograph courtesy of the Sarawak Museum)

In a study of 18 extended burials, Brooks and Brooks (Reference Brooks and Brooks1968) noted they displayed a limited number of standardised arm positions which appeared to correlate with the sex of the individual. An expanded data set of 29 burials (28 adults, young-adults, or sub-adults, and one infant) broadly corroborates this observation (Lloyd-Smith Reference Lloyd-Smith2009, 130–6). Five arm positions (for both the right and left arm/hand) are recognised in burial terminologies (Sprague Reference Sprague2005, 99): extended parallel to the body; lower arm bent so the hand rests on the pelvic area; lower arm bent at 90° across the stomach; lower arm bent up diagonally across the chest area; and finally, where the arm is folded back so the hand rests on the same side shoulder. The number of combinations of right and left arm positions is therefore 25. In the West Mouth Neolithic cemetery 12 positions are present. The majority of extended burials displayed one of two symmetrical arm positions, either with both hands resting on their respective shoulders (Fig. 9; Position G), or crossed over the chest (Position F). Both these position were displayed equally by male and females. Three burials had both arms extended by the sides of the body (Position A): B75a was younger child; B36 was an adult female located at a distance from the main area of the cemetery towards the cave entrance; and B232, an adult male with congenitally missing radii bones resulting in malformed and shortened forearms (Zuraina Majid et al. Reference Zuraina, Lim, Arif, Samsuddin, Nizam, Zafarina and Norazmi2005). By pooling those positions (seven burials) where one hand rests over the pelvis into a single ‘hand-on-pelvis’ position, and another (five burials) where one arm is folded at 90° over the body, the number of ‘standardised’ positions, is reduced to six (Table 2). The ‘hand-on-pelvis’ position is displayed by five females, one sex unknown, and one male (B126) although here curiously holding a charred and notched stick (Brooks & Brooks Reference Brooks and Brooks1968, 70; B. Harrisson Reference Harrisson1967). The second non-symmetrical position, where the lower right arm is at 90° across the body, is displayed by two adult males and two adults with unknown sex.

Fig. 9

Arm positions displayed by extended burials in the West Mouth, Niah Cave (illustration: L. Lloyd-Smith)

Table 2

Standard arm positions displayed by extended burials in the west mouth, niah cave

Higher order variables concern the application of red pigmentation and associated material culture, including the burial container (wooden coffins, bamboo matting [sometimes identifiable as completely wrapping the body], and ceramic jars; Fig. 10), and other grave goods (stone pounders and adzes, shell or bone ornaments [Fig. 6], wooden sticks, and organic beads). However, the West Mouth Neolithic burials are not individually rich in grave goods and there is no evidence for significant social differentiation represented in material artefacts. Only 16 of 66 initially investigated extended burials, and 16 of 59 secondary burials were recorded with portable grave goods (B. Harrisson Reference Harrisson1967). It can be noted, though, that cuprous metal is recorded with two burials, including what is described as a ‘bronze knife’ found on the chest area of extended burial B36 (ibid., 159). A radiocarbon date of 3182 ± 26 bp or 1501–1413 cal. bc (OxA-18562) obtained on a sample of coffin wood from B36 probably reflects old wood. However, in 2007 a fragment of a bronze clasp(?) was recovered from the cremated remains of burial B18 (Fig. 11), from which a sample of cremated bone produced a date of 2577 ± 33 bp or 812–564 cal. bc (OxA-22058); the oldest secure date for metal in Island South-east Asia. The bronze artefacts probably originated from mainland South-east Asia, and along with the ling-ling-o pendant, demonstrate that the Niah community, though not spectacularly rich, was part of a regional network around the South China Sea along which material goods, people, and ideas circulated from at least the early 1st millennium bc.

Fig. 10

Unburnt secondary burial B159 in a large earthenware jar, during excavation in 1965. Burial B159 is dated to 1195–977 cal. bc (photograph courtesy of the Sarawak Museum)

Fig. 11

Bronze clasp or hook found amongst cremated remains of burial B18, dated to 812–564 cal. bc (photograph: L. Lloyd-Smith)

Burial Groups

By visual inspection the West Mouth burials can be divided into a number of spatially discrete groups (Table 3; Fig. 12). The rationale for dividing the site thus was as a working tool to explore spatial and stratigraphic patterning in mortuary practice and as a basis for a radiocarbon dating programme to investigate ideas concerning the growth and history of the cemetery. The dating results refined, but did not substantially alter, the integrity of, these spatial groups, supporting the interpretation that the groups are ‘real’ and the product of deliberate and meaningful organisation of the cave for burial. Twenty-five burials (16 flexed, two flexed/decapitated, three seated, and four secondary burials) that form Burial Groups 1–4 date to the Early Holocene (Lloyd-Smith Reference Lloyd-Smith2013) and are not included.

Table 3

Spatial burial groups in the west mouth, niah cave

Depths refer to the level at which the burial deposit was first exposed below the cave surface. Burial Groups 1–4 date to the Early Holocene (Reference Lloyd-SmithLloyd-Smith 2013)

Fig. 12

Proposed Neolithic and Post-Neolithic spatial burial groups; Burial Groups 5, 6 ∼ Post-Neolithic; Burial Groups 7–12 ∼ Neolithic; 15 darker indicated burials (four comprising two overlapping burials) are located in isolated positions between the defined burial groups (illustration: L. Lloyd-Smith)

Of the 182 burials that date after c. 1500 bc, the vast majority can be grouped into five discrete groups within the main cemetery area: a central group (Burial Group 7), surrounded at short distances to the north-east, north, north-west, and west, by four satellite groups (Burials Groups 8–11). Beyond these, lie three smaller groups all with shallow, sub-surface burials: Burial Groups 5 and 6 extend from the western edge of the main cemetery; Burial Group 12 comprises three secondary burials tucked under a large overhang in the north-west corner of the cave. There are also 16, more or less, isolated burials.

Radiocarbon Dating

Absolute dating of the later prehistoric burials in the West Mouth, Niah Cave, is problematic. Seven dates from the 1960s on coffin wood (four samples), bamboo matting (one sample), and human bone (two samples) have wide error margins and fall between 2150 cal. bc and cal. ad 600 (B. Harrisson Reference Harrisson1967; 1968). A subsequent series of 20 dates on collagen from human bone also have wide error margins spanning 4000 cal. bc to cal. ad 900 (Brooks et al. Reference Brooks, Helgar and Brooks1977) and are now also regarded as unreliable due to the, then unrecognised, problem of variable bone diagenesis in tropical environments (Spriggs Reference Spriggs1989).

The recent dating programme was also not without problems. Of 27 submitted bamboo samples, 12 failed to produce a result, while eight produced dates that are spurious. What appeared in the field to be preserved bamboo often was only silica and contained insufficient carbon to date. Furthermore, bamboo samples from which dates were obtained, often produced dates which were older than expected from the stratigraphic sequence of burials (securely dated using cremated bone samples). Similarly, of six wood samples taken in 2006 from exposed in situ coffins, three failed to produce a date; two produced modern dates (post-ad 1950); and one a date in the last 500 years (there is no reason to suspect that the any of West Mouth burials belong to the developed Metal Age, as is the case with other Niah Cave sites where burials are found with abundant Chinese ceramics; B. Harrisson Reference Harrisson1958; T. Harrisson Reference Harrisson1959). These results highlight an apparent problem in the diagensis of organic plant material since initial exposure in the 1960s. The most secure of the recently dated samples from burials are those obtained on cremated human bone. In total, including all samples which produced a result, there are 44 dates from 38 burials (Table 4).

Table 4

Radiocarbon dates from 38 neolithic burials in the west mouth, niah

Burial types follow Lloyd-Smith (Reference Lloyd-Smith2011). Dates marked with asterisks published by B. Harrisson (Reference Harrisson1967, 1968). Eval = date evaluation: tick, accepted with caution; cross, rejected

Radiocarbon dates on charcoal indicate a gap in the use of the cave between c. 6000 and c. 2000 bc (Krigbaum Reference Krigbaum2001; Lloyd-Smith Reference Lloyd-Smith2009). The first post-gap charcoal date has a wide error margin, calibrating to 2050–1100 bc (3285 ± 168 bp; AA-27957; Krigbaum Reference Krigbaum2001, 69). It is not until after 1500 bc that there more accurate and abundant radiocarbon dates on stratified charcoal. Before this date, any human presence at the site was ephemeral and sporadic.

In evaluating each burial date, excluded are those with excessive error margins from burials: B60C (GX-0721), B75 (GrN-7204), and B159 (GX-14280). Four dates produced on bamboo samples from burials B41 (sample dated twice), B216, and B134 are also excluded as probably too early due to the apparent problem of bamboo diagensis. The date from burial B36 found with a ‘bronze knife’ is almost certainly the product of old wood. Likewise the date from burial B211 is rejected; B211 was parallel to extended burial B178A which overlay flexed burial B205, found with a polished adze and from which a date of 2986 ± 29 bp was obtained. The earliest radiocarbon date of 4216 ± 29 bp from burial B185 is also thought unreliable. As well as being over 1000 years earlier than all other dates from the cemetery, the sample was a small loose piece of wood from the grave fill. Also rejected are those dates produced on wood and bamboo which date to the last 2000 years. As will be discussed below, there was certainly activity in the West Mouth during the early centuries ad, with which the in situ wooden stake in B217, dated to 1567 ± 27 bp, may be associated, but it is not thought to date the burial itself which is in the centre of the cemetery and is surrounded by burials and a coherent series of dates between 1000 and 200 bc. The recent dates from burials B138, B133, and B60A, along with the modern dates (post-ad 1950) from burials B113, B138, and B164 are rejected as probably the result of plant diagensis since initial exposure.

There remain 27 radiocarbon dates on samples of bamboo, wood, and cremated bone from 25 burials that are cautiously accepted to build a conservative but secure absolute chronology for the West Mouth Neolithic cemetery (Fig. 13). It is proposed that the cemetery dates between c. 1500 and c. 200 bc, during which six phases of burial can be discerned: primary flexed (1500–1300 bc), primary extended (1300–900 bc), unburnt secondary burial (1000–700 bc), secondary burial cremation (800–500 bc), followed by a second phase of primary extended burial and final a second phase of unburnt secondary burial (500–200 bc). While such a neat and uniform history of burial practice probably did not exist, this chronology is proposed as representing the major shifts in normative burial practice that took place during the course of the Neolithic c. 1500–200 bc.

Fig. 13

Proposed radiocarbon chronology for the Neolithic cemetery

The Neolithic Cemetery Chronology

Phase 1: Flexed burial, c. 1500–1300 bc

The use of the West Mouth for burial by people with Neolithic material culture began with flexed burials dating to sometime between c. 1500 and c. 1300 bc. Based on stratigraphic evidence, the earliest Neolithic burial dated is B205 (Fig. 14). This was in a tightly flexed position, contracted with knees against the chest, like most of the Early Holocene burials, but was accompanied by a lenticular polished stone axe, a classic item of Neolithic material culture. Compressed plant remains overlying B205, possibly associated with a later extended burial (B178A, Burial Group 9), were dated to 2986 ± 29 bp or 1371–1123 cal. bc (OxA-13491).

Fig. 14

Early Neolithic flexed burial B205. Burial B205 was originally discovered at the close of the 1967 excavations and left exposed, after which it became covered over by the accumulation of fresh guano. It was rediscovered in a trampled condition in 2003 during fieldwork of the NCP. Left: re-exposed burial in 2003 looking NE; scale 1 m. Lower right: burial as the skeletal elements were being lifted, at which time the polished adze was discovered on the left chest area with the cutting edge facing towards the feet, looking N; scale 10 cm with 5 mm divisions. Top right: the polished lenticular adze; scale 10 cm with 5 mm divisions (photographs: G. Barker)

The tightly flexed burial position and the inclusion of a complete polished stone adze distinguish B205 from all other primary burials in the Neolithic cemetery. The adze places this burial in the general Neolithic period and thus after the mid-Holocene gap. The earliest possible date for B205 is c. 3000 bc, the lower limit of the radiocarbon date from B185 discussed above. A more conservative estimate based on the abundance of charcoal dates, would place this burial to between 1500 and 1300 bc. On current evidence it appears that this initial period of Neolithic burial was sparse and sporadic. However, it is very likely that further flexed burials remain undiscovered at the site, in particular underneath the centre of the cemetery where the majority of extended burials remain exposed and in situ.

Phase 2: Primary extended burial I, c. 1300–900 bc

There was a major change in mortuary practice after c. 1300 bc: a formalised and structured cemetery was laid out with rows of extended burials, frequently in wooden coffins or bamboo matting. The establishment of the cemetery appears to have been decisive and once underway expanded rapidly. The earliest secure date for extended burial is c. 1400–c. 1200 bc. Dates which fall in this range are from extended burials B75a, B192, B184, and B102 in Burial Group 7; from plant remains overlying flexed burial B205 under Burial Group 9; and from extended burial B46 in Burial Group 11. These secure dates indicate that extended burial was practised contemporaneously in Burial Groups 7, 9, and 11 (Fig. 15). The extended burials in Burial Group 8 also probably date to this phase, while those in Burial Group 10 date to the fifth phase of the cemetery (see Phase 4).

Fig. 15

Phases 1 & 2: Early Neolithic burials with burial numbers discussed in the text indicated. Phase 1 flexed burial B205 is located underneath burial B178A. Phase 2 comprised rows of primary extended burials (illustration: L. Lloyd-Smith)

Burial Group 7 occupies a topographically central location from where it is possible to see all other areas in the front of the cave mouth as well as far into the northern and southern passages. Very few of the extended burials overlap or truncate one another and it is reasonable to assume that the cemetery displays a degree of horizontal stratigraphy. Burial Group 7 is by far the largest group in the cemetery, and it follows that it was added to most often and probably in use over the longest period of time. While the radiocarbon dates do not support the interpretation that Burial Group 7 precedes any of the other groups, it was certainly the main focus once the cemetery was established. The earliest accepted date from near the centre of Burial Group 7 is from B192, from which a sample of bamboo produced a date of 2950 ± 60 bp or 1379–1001 cal. bc (OxA-18393). A sample of wood from B219, also in the centre of Burial Group 7, produced a date of 2847 ± 27 bp or 1114–924 cal. bc (OxA-14028).

The extended burials in Burial Group 7 are aligned along six north–south rows. As these grew in length, the overall shape of the cemetery's central area took form. Once the spaces between each row were filled in, burial spread to the periphery, by which time orientation shifted to reflect the oval shape of the combined cluster of burials. This is particularly noticeable on the north-eastern edge of Burial Group 7. Two inferences follow from this interpretation: first, that each row may represent a social group; and second, that these social groups (in Burial Group 7) were more closely associated with each other than with the other burial rows, in Burial Groups 8, 9, or 11.

Away from the densely packed central area, it is possible to identify pairs of extended burials: B2/B11, B3/B191, and B178A/B211 in Burial Group 9; and B47/B48 in Burial Group 11. Isolated burial pairs B22/B24 and B35/B36 should also be mentioned here. In the case of Burial Group 11, the row of extended burials appears to focus on, and originate from, the male and female burial pair B47/B48, where the burials even appear to be positioned in the grave so that their bodies were turned to face one another (Fig. 16). Burials B47 and B48 also displayed complementary arm positions. Demographic and arm position data in other burial pairs are patchy, but in the case of B3/B191 we again see an adult male and female. On the basis of these observations it is suggested that the rows of extended burials which over time evolved into the distinct Burial Groups 7, 8, 9, and 11 are likely to represent family lineages, households, or perhaps settlement groups.

Fig. 16

Primary extended burials B47 and B48. Burial numbers added to original photograph (photograph courtesy of the Sarawak Museum)

The arm positions of extended burials tentatively correlate with isotopic and demographic data. The burials with both arms laid straight by the sides of the body (Position A) are, in differing ways, abnormal: B36 was an adult female identified as non-local to the Niah area (Valentine et al. Reference Valentine, Kamenov and Krigbaum2008); B75a was a rare infant extended burial; and B232 was an adult with congenitally malformed lower arms (Zuraina Majid et al. Reference Zuraina, Lim, Arif, Samsuddin, Nizam, Zafarina and Norazmi2005). One possibility is that the other arm positions represented some form of social identifier for (‘married’?) adults belonging to the local population. Infants were not assigned an arm position ‘identity’ in burial; the B36 woman clearly did not have such an identity, even though she had moved to, and died at, Niah; and though the individual with malformed arms was cared for into adulthood, their physical disability appears to have negated the positioning of the arms in burial (whether they were ‘married’ or not). The counter to this interpretation is burial B3: an adult male also identified as a non-local individual (Valentine et al. Reference Valentine, Kamenov and Krigbaum2008), buried with the right hand resting on the same side shoulder and left arm bent up across the chest. There is, though, a strong likelihood that B36 was a completely isolated burial. Except being listed in the published original classification, no field records, photographs, or skeletal remains exist for ‘adjacent’ burial B35 (B. Harrisson Reference Harrisson1967), indicating it may in fact simply be a duplicate number. In which case, the other arm positions may indeed display social identities, and B36's ‘non-positioned’ arms highlight the singular nature of this individual: non-local, buried with a bronze knife, and in an isolated location separate from all burial groups.

The most common position, used for both sexes, was with the hands resting on or near their respective shoulders (Position G). The second most common, also used for both males and females, was with the arms crossed over the chest (Position F). The distinctive and pre-dominantly female hands-on-pelvis burials were largely restricted to the eastern edge of Burial Group 7 and the adjacent Burial Group 8, whereas the male burials with one arm at 90° were widely distributed in the cemetery. If, as proposed, the rows of extended burials represent lineages of family descent, the contrasting distributions of the female and male burials indicated by the two asymmetric arm positions, could imply that the Neolithic societies of Niah were organised according to a partial system of matrilocal post-marital residency, whereby the husband moved to the wife's family group or household, with married couples then being buried together, close to the group representing the female line of descent. While the strontium isotope data make it clear, though, that on occasion non-local women (B36) were also joining the community, and her isolated position in burial (with or without B35) suggests that she was not fully incorporated into any of the ‘local’ groups.

Phase 3: Secondary unburnt burial, c. 1000–700 bc

In all areas of the cemetery, except Burial Group 10, secondary burials overlay extended burials (Fig. 17). In many instances there existed a direct spatial correlation between unburnt secondary burials and extended primary burials, which are interpreted as representing historically-linked burial events: that is to say, the grave of the extended burials was re-opened by first or second generational descendants of those interred in the primary burial. In the largest central group, Burial Group 7, many of these re-opened burials are located around the periphery, supporting the hypothesis for a degree of horizontal stratigraphy in the growth of the cemetery and that these extended burials belong to the end of the initial phase of primary burial.

Fig. 17

Phase 3: First phase of unburnt secondary burial in the Neolithic cemetery. Burial numbers discussed in the text are indicated. Phase 3 burials indicated by darker shading, overlaying lighter shaded Phase 2 primary burials (illustration: L. Lloyd-Smith)

However, on current data it is not these primary/secondary burial sequences in which the earliest radiocarbon dates for secondary burial occur. Rather, the earliest reliable date for secondary burial is from the largest jar burial B159 (Burial Group 11), from which a sample of burnt wood (a branch with a short internal age) produced a date of 2889 ± 29 bp or 1195–977 cal. bc (OxA-20995). Burial Group 11 is further distinguished by containing the highest (equal) ratio of adults given secondary burial (five) compared with those given primary burials (five) of all burial groups, and by the fact that none of the secondary burials was cremated. In contrast, cremation burials are found in all other burial groups in the main cemetery, all producing radiocarbon dates that calibrate to between 812 and 521 bc. In this light, the early date from secondary burial B159 indicates that Burial Group 11 had ceased prior to the emergence of the practice of cremation.

A focus on skulls is seen in a number of extended burials with exhumed (missing or replaced) skulls, including B139, B192, and B211 (Fig. 18), and in a number of secondary burials of intact or partial crania, including B45b, B75b, B119, B153, and B186. Many of the secondary burials, both unburnt and cremated, were lacking cranial remains, suggesting the retention (and curation?) of skulls, while the rest of the skeletal remains were buried in the cemetery. The limited number of extended burials with skulls removed indicates that this change in burial ritual was restricted to certain members of society. In the case of B211 the pit excavated to retrieve the skull was backfilled with a substantial quantity of wild pig bones (Barker et al. Reference Barker, Barton, Bird, Cole, Daly, Gilbertson, Hunt, Krigbaum, Lampert, Lewis, Lloyd-Smith, Manser, Menotti, Paz, Piper, Pyatt, Rabett, Reynolds, Stephens, Thompson, Trickett and Whittaker2002b, 118) suggesting that feasting or offerings were part of these secondary rituals.

Fig. 18

Extended burial B211 with skull removed by the excavation of a pit back-filled with wild pig bones (excavated). Scale: 1 m; 20 cm divisions (photograph: G. Barker)

That the skulls exhumed from primary burials, and possibly those retained from the initial secondary burial deposition, were subsequently (re)buried, indicates that the curation the skulls was to serve a particular purpose and at some point in time this function was completed. Alternatively, it might have been that the (re)burial of the skulls was a significant part of the ritual process. The exhumation and retention of skulls suggests that ancestors played an important role in living society. The retrieval and reburial of skulls from extended burials supports the idea that the transition from primary to secondary burial was an historical process. It appears that the transition to secondary burial originated in the practice of retrieving the crania of certain individuals given primary burial. The stratigraphic relationships between primary and secondary burial are the same in the four oldest burial groups (7, 8, 9, and 11), suggesting that the widespread adoption of secondary burial constitutes a distinct phase in burial practice. The implication is that the widespread transition to secondary burial was rapid and occurred within the space of one or two generations.

Phase 4: Secondary burial cremations, c. 800–500 bc

The medial radiocarbon dates produced on eight cremated bone samples varied by only 79 years, and because of a plateau in the calibration curve they produced very similar calibrated dates, all falling within the range 812–521 cal. bc. Cremation burials are located in all central burial groups except Burial Group 11 which probably had ceased prior to the transition to cremation (Fig. 19). Those in the Burial Group 7 form a discrete cluster, where they are spatially independent of the early primary or unburnt secondary burials. Moreover, Burial Group 10 appears to have been initiated by cremation burial. Such a strong spatial discontinuity in cemetery organisation hints at a chronological hiatus in burial activity, with the cremation burials representing the subsequent re-use of the cemetery. Unfortunately, due to the sharp break of slope in the radiocarbon calibration curve at 800 bc, it is difficult to investigate this possibility further.

Fig. 19

Phase 4: Secondary burial cremations and two multiple burials. Burial numbers discussed in the text are indicated. Burials including cremated remains indicated with darker shading. Phase 1–3 burials, lighter shading (illustration: L. Lloyd-Smith)

The practice of cremation forms the second phase, and elaboration, of secondary burial at the site. All cremated remains display green-fracturing. The dead were cremated soon after death, rather than the burning of dry skeletal remains at a later date. All the cremated remains from 24 secondary burials in the Neolithic cemetery were also stained red, presumably ochre. This compares with only four instances where red coloration was recorded in the total of 41 unburnt secondary burials. The interior and articular surfaces of cremated bones were coated in a layer of red pigmentation, indicating that the colour was applied to the bones in liquid form. Cremations also have a stronger association with earthenware burial jars than unburnt secondary burials. Of the 24 burials containing cremated remains, seven were in ceramic vessels, compared with five of the 41 unburnt secondary burials (Lloyd-Smith & Cole Reference Lloyd-Smith and Cole2010). ‘Multiple’ burial B60B-D included a discrete cremation deposit on the base of the over-size log coffin, suggesting that this special burial and multiple burial B138 in Burial Group 8 belong to Phase 4 as well.

Phases 5/6: Primary extended burial II/secondary burial II, c. 500–200 bc

Varied burial activity continued at several different locations within the cemetery after 500 bc (Fig. 20). In Burial Group 10, funerary practice switched back to primary extended burial. A piece of bamboo from extended burial B50 produced a date of 2412 ± 50 bp or 750–397 cal. bc (OxA-18400); B50 overlay two jar burials, B85 and B96 both of which contained cremated remains. Further out from the cave wall, a fragment of coffin wood from B21 produced a date of 2317 ± 30 bp or 414–233 cal. bc (OxA-18453).

Fig. 20

Phases 5 & 6: Second phases of primary extended and unburnt secondary burial. Burial numbers discussed in the text indicated. Phase 5 & 6 burials indicated with darker shading. Phase 1–4 burials, lighter shading (illustration: L. Lloyd-Smith)

Unburnt secondary burial occurs elsewhere including a sequence of stacked jar burials B221 and B190 placed in the approximate centre of Burial Group 7 (Lloyd-Smith & Cole Reference Lloyd-Smith and Cole2010). A piece of burnt wood, with short internal age, from inside B190 produced a date of 2308 ± 35 bp or 477–209 cal. bc (OxA-11548). Burials B221/B190 appear to be situated on their own and placed at the centre of the cemetery, where the burials in the immediate vicinity date to at least 400 years earlier. The location of the stacked burial jars represents a renewed focus upon what would by that time have been an extremely ancient place of burial.

Elsewhere, we see the beginning of a shift away from the main cemetery area and towards the front of the cave, along the cave wall. Burial Group 12 comprises three sub-surface unburnt secondary burials (B29, B30, and B78) deposited just under the southern edge of the overhang. They have strong ceramic associations, including (possibly re-used) double-spout fragments (Franca Cole, pers. comm.). There are no secure absolute dates for these burials, but the δ¹³C isotopic data from B30 are comparable with the majority of the extended and secondary burials in the main cemetery (Krigbaum Reference Krigbaum2001, 181). It is proposed that Burial Group 12 represents an expansion along the cave wall during the second half of the cemetery's use sometime between 800 and 200 bc. Here also can possibly be included two isolated jar burials, B233 into the cave passage (Fig. 3), and B154, along with secondary burial B37 (Fig. 20).

The latest radiocarbon date for burial within the Neolithic cemetery is from unburnt secondary burial B103. A sample of burnt wood (a branch with a short internal age) produced a date of 2223 ± 28 bp or 383–203 cal. bc (OxA-18451). As in the case of the stacked burial jars B221/B190, this burial is surrounded by much earlier burials. B103 overlay the western end of two extended burials B102 and B104. A sample of bamboo from the burial byre under B102 produced a date of 2885 ± 70 bp or 1289–899 cal. bc (OxA-18396), at least 500 years prior to B103. Secondary burial B103 was covered by a variety of earthenware sherds, suggesting the deliberate re-use and movement of pottery from elsewhere in the cemetery (Franca Cole, pers. comm.). Such practices would have maintained a direct material link between the members of the funerary party and their ancestors who had been buried at the site for over a thousand years. What distinguishes B103 from the earlier Neolithic burials, though, is a markedly different δ¹³C isotopic signature. After a thousand years of δ¹³C values interpreted as representing open-canopy, probably agricultural, subsistence practices (Krigbaum Reference Krigbaum2001, 193), B103 has a value comparable to those of the Early Holocene flexed burials discovered at the front of the cave and under the overhang, possibly indicating the beginning of a shift back to a more varied and mixed subsistence.

Post-Neolithic Burial

A cluster of sub-surface features towards the front of the cave, originally classed as ‘disturbed/doubtful remains’ (B. Harrisson Reference Harrisson1967, 190), have been reclassified as (disturbed) loosely flexed burials and unburnt secondary burials, including skull burial. Together these burials comprise Burial Group 6 (Fig. 21). The disorganised appearance of Burial Group 6 contrasts with the formalised organisation of the Neolithic cemetery. This and the conspicuous distance between Burial Group 6 and the Neolithic cemetery suggest that the two are not contemporary. Neither do flexed burials B32a or B34 in Burial Group 6 resemble flexed burial B205, which represents the earliest phase of Neolithic burial at the site.

Fig. 21

Proposed Post-Neolithic Burial Groups 5 & 6. Burials discussed in the text indicated, Neolithic burials, lighter shading (illustration: L. Lloyd-Smith)

Adjacent to Burial Group 6 and spreading c. 10 m into the cave were the four semi-articulated secondary burials, which comprise Burial Group 5. These burials contained completely disarticulated skeletal remains mixed with, and overlying, articulated sections of a spinal column (see Fig. 7). As such they appear to represent a foreshortened secondary burial rite, whereby the body was retrieved for final burial before the corpse was fully reduced to bones. One possibility is that these features are associated with secondary burials that were given final burial in one of the burial groups in the main cemetery. If this was the case, however, one would expect many more semi-articulated burials. Furthermore, in three of these burials substantial quantities of disarticulated secondary burial remains were included, indicating they represent complete burials in their own right, rather than the residual elements of separate secondary burials placed in the Neolithic cemetery.

The δ¹³C isotope data reveal interesting correlations between Burial Groups 5 and 6. Flexed burial B34 and skull burial B33 (Burial Group 6), and semi-articulated burial B153 (Burial Group 5), all displayed some of the most negative δ¹³C values of all burials at the site, strongly indicative of closed-canopy environments and hunter-gatherer subsistence practices (Krigbaum Reference Krigbaum2001, 193).

There are no absolute dates available for Burial Groups 5 or 6, but their shallow depths and mixture of earthenware pottery within and under the burial deposits indicate that they date to the Neolithic or later. The spatial separation of these two burial groups from the main cemetery, the disorganised nature of Burial Group 6 compared with the formalised character of the Neolithic cemetery, and the interpretation that the semi-articulated remains represent on-site transformational burial rites, all suggest that the sub-surface burial activity towards the front of the cave mouth was not contemporaneous with the Neolithic cemetery. The spatial and isotopic similarities of these two burial groups, suggests that they constitute a single phase of burial that is subsequent to the main Neolithic cemetery: what I propose to term the ‘Post-Neolithic’. Here also can possibly be included the sub-surface isolated secondary burial B95 at the very front of the cave mouth.

There are, though, similarities between the practice of secondary burial in the Neolithic cemetery and Post-Neolithic burial. A focus on the skull persists and the spatial association of substantial quantities of pig remains found around the semi-flexed and secondary burials in Burial Group 6 are reminiscent of burials in the Neolithic cemetery. We may also see an overlap, or precursor, to the Post-Neolithic closed-canopy isotopic signatures in the strongly negative δ¹³C value of B103. While the upper limit of the date of 200 bc from B103 is taken as the earliest possible date for the Post-Neolithic burial phase, the spatial separation between the Neolithic cemetery and Post-Neolithic burial at the site suggests that a chronological hiatus separates the two phases.