Clarifying ambiguity: a reassessment of the multi-disciplinary evidence for a 10,000 year-old palaeosurface at Kuk Swamp, highlands of Papua New Guinea

Abstract

Shifting cultivation is a widespread agricultural practice in the wet tropical forests of the world, as well as other environments. Short-term cultivation followed by long-term fallow is a human adaptation to the low fertility of soils under tropical forests. Forms of shifting cultivation are likely to be ancient and among the earliest agricultural practices. Here, a multi-disciplinary method for the investigation of shifting cultivation in the past is outlined. The method is grounded in direct archaeological evidence and palaeo-pedological evidence for soil formation processes associated with cultivation practices, which are interwoven with contextual palaeoecological and geomorphological evidence for landscape transformation. Hence, these multi-disciplinary lines of evidence are recast to assess whether a short period of use of a c. 10,000 year-old palaeosurface at Kuk Swamp in montane Papua New Guinea represents shifting cultivation or alternative types of plant exploitation in the past.

Introduction

The highlands of New Guinea have archaeological evidence for some of the oldest agricultural practices in the world. Remains of former cultivation plots have been excavated at several wetlands in the highland interior, principally at the site of Kuk Swamp in Papua New Guinea (Golson et al., 2017). The earliest evidence for cultivation using mounds dates to c. 7,000–6,400 cal BP, with subsequent ditched field systems dating from c. 4,400–4,000 cal BP (Denham et al., 2003). The significance of earlier archaeological evidence of a short-lived palaeosurface dating to c. 10,000 cal BP on the wetland margin at Kuk is debated; it is uncertain whether it represents a nascent form of shifting cultivation or an alternative form of plant exploitation (Denham, 2018a). Despite this uncertainty, shifting cultivation was potentially practiced on the floor of the Upper Wahgi Valley from c. 10,000–7,000 cal BP (Denham and Haberle, 2008; Haberle et al., 2012). The global significance of Kuk is demonstrated through its inscription on UNESCO's World Heritage List in 2008 (Muke et al., 2007).

The investigation of ancient shifting cultivation in wet tropical forest environments is often problematic. Archaeological evidence of practices associated with plot clearance, cultivation and harvesting do not ordinarily preserve in wet tropical forests where soils have high rates of erosion, pedoturbation and weathering, nor does archaeobotanical evidence for many of the vegetatively propagated crops that people potentially grew, especially roots and tubers. Consequently, shifting cultivation in the past has usually been inferred from changing proportions of disturbance, secondary forest and primary forest taxa, as well as charcoal, in pollen diagrams extracted from wetlands, such as lakes and swamps (Flenley, 1988; Haberle et al., 2012).

Here, a multi-disciplinary method is applied to reassess the multidisciplinary evidence associated with a c. 10,000 year-old palaeosurface at Kuk Swamp. The method weaves more direct lines of evidence from archaeology and sedimentology together with more indirect lines of evidence from geomorphology and palaeoecology to assess whether they plausibly represent shifting cultivation in the past. The method is exemplified with respect to forms of plant exploitation in the highlands of Papua New Guinea, where it was originally developed and applied (Denham, 2007, 2018a, b). The Kuk case study is intended to explore the method's utility to the investigation of cultivation practices based on vegetative propagation in other wet, tropical and subtropical environments where quantifiable archaeobotanical evidence of the plants grown is often lacking.

Agricultural practices in Papua New Guinea

Various forms of agriculture are practiced across Papua New Guinea today and in the recent past (Bourke and Harwood, 2009; Brookfield and Hart, 1971; Powell et al., 1975). These can be classified in terms of intensity and continuity of land use, namely, labour inputs per unit area of land through time (Bourke, 2001). Today, labour inputs are increasingly supplemented by monetary inputs in the forms of fuel, fertilisers and pesticides, which also require consideration in terms of how intensity of land use can be assessed. Traditionally, higher intensity, semi-permanent forms of cultivation occurred in house gardens and in cultivated plots in grasslands and wetlands, while lower intensity forms of shifting cultivation and arboreal exploitation occurred in forests.

House gardens

House gardens are relatively ubiquitous and are located within and adjacent to settlements (Bourke and Allen, 2009: 262, 263). They usually contain a diverse mix of: minor crops, such as greens, herbs, medicines and flavourings; snack foods, including bananas (Musa cvs) and sugarcane (Saccharum officinarum); and, ornamentals (Figure 1). House gardens are tended on a daily basis and are usually cultivated for as long as a settlement is occupied, which may span from a couple of years to decades. Within settlements, old garden areas may be repurposed for building houses and abandoned house sites may be cleared for use as house gardens.

Figure 1

Mounds and raised beds

Semi-permanent forms of cultivation using mounds (of varying size) and raised beds (of varying shape) usually occur in anthropic grasslands, namely where people have caused the degradation of forest environments (Bourke and Allen, 2009: 251–255). Environmental degradation can result from the replacement of forest by grasses with the shortening of fallow cycles under shifting cultivation. Over time, forest species are unable to regenerate and become replaced by secondary forest taxa and eventually by grasses (Haberle et al., 2012; Hope and Haberle, 2005). Periodic burning of the landscape and climatic stressing may increase the rate of forest degradation to grassland. The investment of labour required to clear and till dense root mats in grasslands is offset by the long-periods of cultivation enabled by mounds and raised beds. Today, in many places, the breaking up of root mats in grasslands is initially undertaken by foraging pigs (Sus scrofa) prior to tillage and planting.

Plots can be cultivated for extended periods using mounds and raised beds, although the earthworks are remade after harvesting (Powell et al., 1975: 11). Fertility is enhanced through composting, the incorporation of ash, and crop rotations using legumes. While a plot is in use, the ground may be left fallow for a few months between harvesting and replanting, but once abandoned an extended fallow period of 30–40 years may occur (Powell et al., 1975: 11). Plots cultivated using mounds and raised beds can be planted with a single crop, especially of sweet potato, or inter-cropped with a mix of root crops and vegetables (Figure 2).

Figure 2

Semi-permanent forms of cultivation occur on the floors and lower slopes of the main highland valleys, some of which were degraded to grassland thousands of years ago (Haberle, 2007). It is probably not coincidental that the earliest, unequivocal archaeological evidence for mound cultivation dates to c. 7,000–6,400 cal BP on the wetland margin at Kuk Swamp and directly follows the degradation of the local environment to grassland (Denham, 2003a, 2018a; Denham et al., 2003).

Tree-fallowing occurs in some parts of the highlands, where abandoned plots near settlements and in grassland are deliberately planted with she-oak (Casuarina oligodon), a fast-growing and nutrient-fixing tree. The practice is highly localised to regions of the highland interior (Bourke and Allen, 2009: 245–247), although it likely dates back to at least 1,200 years ago (Haberle, 2007; Sniderman et al., 2009).

Wetland drainage

The cultivation of plots in drained wetlands is widespread in the highlands (Ballard, 1995; Ballard et al., 2013). Individual plots may be cultivated for extended periods and on a long-fallow basis using mounds and raised beds. Wetlands provide relatively nutrient rich and well-watered environments for more continuous cultivation, in contrast to the more highly weathered soils on valley slopes (Denham and Grono, 2017). Ditched field systems can be constructed and maintained by relatively small groups of people, while extensive ditch networks across wetlands require co-ordination and co-operation between multiple communities to ensure proper functioning, especially regular maintenance of major water disposal channels (Ballard, 2017).

The earliest ditch networks and the oldest wooden spade in the highlands date to c. 4,400–4,000 cal BP at Kuk and Tambul, respectively (Denham, 2005; Golson et al., 2017). Ancient ditches can be readily traced in the walls of modern drains and in the floors of archaeological trenches in order to reconstruct former ditch networks across space. Most wetlands on the floors of the major intermontane valleys across the highland spine of New Guinea were drained in some way for cultivation in the past, with many displaying evidence for sequential periods of drainage and abandonment (Ballard et al., 2013; Golson, 1982).

Shifting cultivation

In contrast to house gardens and mound/raised bed cultivation in grasslands and wetlands, shifting cultivation is a relatively low intensity form of agriculture. Shifting cultivation is usually practiced in areas of the fringe highlands and lowlands of New Guinea where extensive tracts of forest survive and where population densities are relatively low (Clarke, 1971; Rappaport, 1968). Forms of shifting cultivation can be “slash-and-burn” or “slash-and-mulch”; the latter occurs where rainfall is abundant throughout the year. Characteristically a fenced plot is cultivated for 1–3 years prior to abandonment and left to fallow for at least 15 years (Bourke and Allen, 2009). Fences are constructed to keep pigs out of cultivated plots. Ordinarily plots are planted with a mix of vegetables, root crops, sugarcane and bananas, although sometimes one crop can predominate (Figure 3). The antiquity of shifting cultivation in the interior of New Guinea, like in other parts of the world, is uncertain and much debated, although it has been inferred primarily from palaeoecological and geomorphological evidence for the early Holocene in the Upper Wahgi Valley (c. 10,000–7,000 cal BP; Denham, 2018a; Denham and Haberle, 2008; Hughes et al., 1991).

Figure 3

Perennials can be harvested and propagated from old garden sites for years after abandonment. In many places, claims to land by individuals or families within a community are based on continued use of an area, including the clearing and making of cultivation plots, as well as visitation during fallow.

Exploitation across the landscape

As well as undertaking forms of higher and lower intensity agriculture, people across New Guinea engage in extensive forms of arboreal exploitation across the landscape (after Denham, 2004a, 2018a; Latinis, 2000). People exploit plants in different habitats within the landscape, not just within cultivated plots. In this sense, some landscapes are effectively hortiscapes, where the distinction between plot and forest blurs because both are domesticated places (Denham, 2018a).

Arboreal exploitation encompasses the exploitation of faunal and floral resources within forests. Faunal exploitation includes fishing and hunting, as well as the gathering of eggs (avian and reptilian), invertebrates (such as insects and grubs) and small vertebrates (such as frogs and lizards) for consumption. Floral exploitation includes forms of arboriculture – namely, the planting, tending and harvesting of tree crops – focussed on starch, protein and oil-rich trees, palms and pandans. Notable species in the lowlands (below 1,200 m) include swamp sago (Metroxylon sagu), marita (Pandanus conoideus), canarium (Canarium indicum), coconut (Cocos nucifera), Barringtonia procera and Terminalia spp. Karuka (Pandanus julianettii) is the only significant, traditional tree crop in the highlands today, whereas other species including castanopsis nuts (Castanopsis acuminatissima) and another pandan (Pandanus antaresensis) were likely more important in the past. Tree crops are more important contributors to diet in the lowlands, where a greater number and diversity of useful trees grow. As well as engaging in forms of arboriculture, people across New Guinea exploit a diverse array of plants for food, construction, ornamentation, making tools and weapons, medicine, toxicity and psychoactive properties, and so on (Powell, 1982a).

A repertoire of practices

People have variably drawn on the repertoire of cultivation and plant exploitation practices to create the diversity of subsistence lifeways across Papua New Guinea today (Bourke and Harwood, 2009), as they did in the past (Denham, 2004a, 2018a). In the highlands, communities are heavily dependent on high-intensity, semi-permanent cultivation using mounds and raised beds on the densely populated floors and lower slopes of the main intermontane valleys. These communities have only limited access to tracts of forest for gathering and hunting on the valley walls and in adjacent valleys. By contrast, in the highland fringe and lowlands, most communities engage in extensive forms of shifting cultivation together with considerable investments in arboriculture, fishing, gathering and hunting, as well as house gardens and semi-permanent plots. For some communities, the planting and management of sago groves replaces the need for cultivation in plots.

Cultivated plants

In contrast to most other parts of the world (Barker, 2006), cultivation practices on New Guinea are predominantly vegetative, namely a plant-part is removed from the parent and transplanted in prepared ground. Vegetative reproduction is clonal and constitutes human exploitation of the asexual reproductive capacity that is innate to many plants (Denham et al., 2020). As well as root crops, such as manioc (Manihot esculenta), sweet potato (Ipomoea batatas), aroids (including taro, Colocasia esculenta) and yams (Dioscorea spp.), a wide variety of plant types are cultivated vegetatively in the New Guinea region today, including: palms (swamp sago); pandans, or screw pines (karuka and marita); fruit bearing herbs (bananas); grasses (including sugarcane); and, green leafy vegetables (Abelmoschus manihot and Rungia klossii) (French, 2006). Forms of agriculture based on vegetative propagation are sometimes referred to as vegeculture (Barton and Denham, 2011, 2018; Yen, 1973). Further, many arboricultural practices on the island entail the transplantation of cuttings and germinated seedlings rather than the planting of trees directly using seed.

The major plants of New Guinea agriculture and arboriculture have diverse domestication and dispersal histories. Several plants are indigenous domesticates, which are still predominantly only cultivated on New Guinea and adjacent islands. Some of these have distinctive altitudinal ranges and are ordinarily cultivated in the highlands (regions above 1,200 m) and highland fringes, including karuka (Pandanus julianettii), rungia (Rungia klossii) and highland pitpit (Setaria palmifolia); while others are ordinarily cultivated in the highland fringes and lowlands, including marita (Pandanus conoideus) and lowland pitpit (Saccharum edule), as well as a range of fruit and nut-bearing trees, including species of Barringtonia, Canarium and Terminalia (see French, 2006).

Of these, the domestication of grasses through vegetative propagation on New Guinea is relatively unusual. Several major domesticates of world agriculture are grasses that were domesticated through sexual reproduction using seed, including: rice (Oryza sativa) and millets (Panicum miliaceum and Setaria italica) in East Asia, maize (Zea mays) in MesoAmerica, and a host of cereals in Southwest Asia (Fuller et al., 2014). On New Guinea, highland pitpit (Setaria palmifolia) and lowland pitpit (Saccharum edule) were domesticated as vegetables and at least one precursor to domesticated sugarcane (Saccharum cvs.) likely originated on the island from wild Saccharum robustum (Denham, 2025; Grivet et al., 2004). These grasses are propagated through planting of leafy lateral shoots (highland pitpit), stalk cuttings (lowland pitpit) and cane segments or division of the root stock (sugarcane) (French, 2006). Lowland pitpit is cultivated for its edible inflorescence, or unopened flower, and highland pitpit for thickened young shoots and stem hearts (French, 2006; Powell et al., 1975). Sugarcane is cultivated for its sugar rich pith, which is chewed as a snack food today, although may have been a more important contributor to diet in the past (Daniels and Daniels, 1993).

Two main cultivar groups of pandanus, or screwpine, are cultivated on New Guinea today: karuka in the highlands and marita in the lowlands. These are propagated vegetatively from crowns and crown-cuttings of old trees, as well as using marita suckers (French, 2006). Like many other cultivated plants karuka can reproduce from self-sown seeds, yet people have focussed on the vegetative mode of reproduction to cultivate them. By contrast, marita is known to only reproduce vegetatively, although a potential wild precursor has been identified in western New Guinea (Keim, 2012). Karuka, or highland pandanus, comprises members of the Pandanus brosimos/iwen/julianettii complex, in which there is a likely domestication cline from P. brosimos (wild-type) to P. julianettii (domesticated) (Cook, 1999). Karuka is grown for its caloric, oil and protein rich nuts (Stone, 1982). Marita, or lowland pandanus (Pandanus conoideus) is grown for its energy and oil bearing fruits.

Numerous green leafy vegetables are cultivated and were probably domesticated on New Guinea, including rungia, oenanthe and crucifer spinach (Rorippa sp.). Many of these plants are widely distributed throughout the wet tropics and subtropics, and likely became enmeshed in multiple independent domesticatory relationships, for instance, Commelina diffusa. Many vegetables are vegetatively propagated from cuttings or rooted stems, while some are planted from seed, including crucifer spinach and Amaranthus tricolor (French, 2006).

Further, a host of major crop plants were likely domesticated on New Guinea, or trace part of their domestication histories to the island, including bananas (Musa cvs.; Martin et al., 2023; Perrier et al., 2011), sugarcane (Saccharum cvs.; Denham, 2025; Grivet et al., 2004), some diploid taro cultivars (Colocasia esculenta; Lebot et al., 2004) and some yam species (Dioscorea spp.; Malapa et al., 2005). Over millennia, however, indigenous lineages of some major cultivars have been successively replaced by introduced varieties of the same crop plant. These sequential transformations can be inferred from the marginalisation of indigenous diploid taro cultivars by introduced triploid cultivars of the same crop in large parts of the island today (Kennedy and Clarke, 2004).

The cultivation, domestication and dispersal histories, as well as ecology and phytogeography, of most plants of New Guinea agriculture are poorly understood. For example, kudzu (Pueraria montana var. lobata; Barrau, 1965) and bottle gourd (Lagenaria siceraria; Kistler et al., 2014) are likely to be ancient introductions to the island, but the timings and mechanisms of introduction are unknown. Clearly, several South American domesticates were introduced more recently to the island and were most likely associated with European and Chinese exploration of the Pacific, including sweet potato (Ipomoea batatas), manioc (Manihot esculenta) and maize (Zea mays). Sweet potato and maize were grown extensively in the interior of New Guinea at the time of first exploration by Europeans/Australians during the twentieth century (Ballard et al., 2005; Bulmer, 1982). Sweet potato is today the most significant staple crop on the island and is especially dominant in the highlands for food and fodder; while manioc is becoming increasingly important.

As well as having diverse geodomestication histories, cultivated plants on New Guinea exhibit varying degrees of domestication (Denham, 2004a, 2018a). Although many groups differentiate between plants grown by people and those that grow themselves (Sillitoe, 1983), there is often an overlap and diffuse boundary between plants that are classified as wild and cultivated. There are gradients (Caballero, 2004) or degrees of domestication (Denham, 2018a), in which many plants are best considered to be “semi-domesticated” (Yen, 1985).

Some plants are clearly domesticated, in the sense that they are more reliant on people to propagate them, including triploid (effectively sterile) banana and taro cultivars, effectively sterile greater yam (Dioscorea alata) and asexually reproducing marita. There are also a range of plants that are cultivated intensively and mostly propagated vegetatively from cultivated plants, such as aibika (Abelmoschus manihot), lowland pitpit, sugarcane and winged bean (Psophocarpus tetragonolobus). By contrast, other plants may be cultivated intensively, yet wild, feral and self-sown plants may also be variably exploited and integrated into cultivated plots or stands thereby diluting any domestication traits, such as highland pitpit, karuka and oenanthe (Oenanthe javanica).

Conceptual and methodological differentiation of shifting cultivation in the past

Figure 4

Given the character of agricultural practices and cultivated plants, as well as the nature of the multidisciplinary record for the past, a practice-based method has been developed to identify early agriculture on New Guinea (Denham, 2004a, 2007, 2009, 2011, 2018a; Denham and Haberle, 2008). Archaeological remains of former practices are more fundamental to the identification of early agriculture on New Guinea than other lines of evidence, such as domestication traits in archaeobotanical assemblages of plant macrofossils (seeds, nuts, fruit stones, charcoal and so on) and microfossils (phytoliths, starch grains, parenchyma fragments), or palaeoecological proxies (mostly pollen and microcharcoal records) of environmental change (Denham and Haberle, 2008; Haberle et al., 2017).

Conceptually, practices are ontologically prior because they emplace plants in contexts associated with former cultivation (Denham, 2018a). Methodologically, archaeological artefacts, features and deposits associated with former practices ground archaeobotanical and palaeoecological evidence in specific contexts associated with former cultivation and plant exploitation. These ideas are considered here with respect to the investigation of shifting cultivation at Kuk Swamp in the Upper Wahgi Valley.

Early forms of shifting cultivation are not likely to be clearly differentiable in the multidisciplinary record from other forms of human practice, including the exploitation of naturally formed gaps and the creation of resource-rich patches in tropical forests (Denham, 2015, 2016, 2018a). There are correspondences and continuities in the character of human activities that occur in gaps, patches and plots. However, it may be anticipated that the cumulative effects of these practices through time are different.

Gaps form in tropical forests through tree fall, landslides, lightening strikes and so on. Once a gap, or canopy space, forms it is rapidly colonised by species that often find it hard to flourish under mature forest canopy (Ghazoul and Sheil, 2010). Some of these gap colonisers are favoured by people and animals, such as bananas and yams. As a result, these gaps can become localised foci of plant and faunal exploitation. Gap dynamics – like that of edges and ecotones - form part of ecosystem processes within tropical rainforests and do not necessarily lead to long-term, cumulative and directional changes in forest composition.

People create “gaps” or patches in forests through the cutting down of trees, localised burning and preventing forest regeneration within gaps (Denham and Barton, 2006). Patches are artificially created resource-rich environments within forests, which effectively mimic gaps because they are designed to promote the growth of favoured species and can become foci for hunting and gathering of faunal resources (invertebrates, grubs and so on). If localised, small-scale and short-term, the creation of patches in tropical forest will have limited long-term effects on tropical forest composition; people are merely another element of the ecosystem contributing to gap dynamics (Denham, 2018b). In this scenario, patches have no archaeological visibility and very low palaeoecological visibility, perhaps only being visible through slightly elevated microcharcoal records.

If the density and periodicity of patch creation increases within a given area, they may have deleterious effects on forest ecology, health and composition. At this point, resource intensification shifts from a patch to a landscape-scale. In this situation, there may still be no archaeological record of the patches themselves, although an increased human presence could be registered at occupation sites within the region. Palaeoecological records would increasingly diverge from those anticipated based on prevailing climates with increases in secondary, disturbance and potentially useful taxa, as well as microcharcoal, which are often calibrated with reference to modern analogues (Haberle, 1993; Hope and Haberle, 2005; Powell, 1970).

By contrast, plots are defined areas of ground that have been cleared of most, if not all pre-existing vegetation for planting. Large trees may often be left in situ within plots, sometimes with pruning or pollarding, to provide shade, aid soil retention or for culturally-specific reasons, such as bird-of-paradise (Passeriformes) display trees, useful tree products, association with an ancestor, and so on. Planting, harvesting and working of the ground is usually undertaken with wooden tools, such as dibbles and digging sticks. In forms of shifting cultivation, a plot is cultivated for a year or so, then abandoned and left fallow for an extended period.

If the density and periodicity of plot clearance is low within a given area, there will be limited impacts on forest ecology, although there may be slightly elevated levels of microcharcoal, secondary taxa and disturbance taxa in palaeoecological records for the vicinity. However, if the density and periodicity of cultivation plots, albeit individually small in area, increases within the landscape, there will be degradation of the forest to secondary forest, species indicative of disturbance and gardening, and eventually grassland. These ecological transformations will become visible in palaeoecological records as cumulative and directional changes in landscape ecology (Haberle et al., 2012).

Although the modes of formation and purpose of gaps, patches and plots can be differentiated in the abstract, in practice there is considerable overlap between each in terms of categorical classification and activities undertaken by people. In terms of categories, people may artificially maintain gaps through localised clearance and burning/mulching, thereby creating a patch, or more systematically clear, burn and plant them to create a plot. Conversely, when left to fallow, old plots become patches within the forest that people maintain to harvest and tend useful plants. In terms of the practices engaged in, people may incidentally plant, or “seed” gaps and patches with useful species, as well as systematically plant in plots; while people may also harvest and sometimes propagate useful plants from gaps and patches for planting in plots. Basically, people engage in the same range of practices within gaps, patches and plots - including clearing vegetation, burning/mulching, planting, tending and weeding, harvesting (including digging) and so on; however, there are slightly different combinations and emphases within each.

Following these considerations, there are methodological problems in seeking to differentiate human exploitation of gaps and patches from plots used for shifting cultivation in the past. The archaeological traces of each type of exploitation are faint and likely to be poorly preserved in tropical soils under forest. Archaeological and palaeo-pedological evidence for digging, potentially associated with planting or the harvesting of tuberous plants, does not usually preserve in most contexts due to subsequent soil formation, later cultivation practices, erosion and deep weathering profiles. The significance of the occasional features that are preserved beneath the canopy will often be unclear. Exceptions include wetlands where evidence can be preserved following inundation and burial of old palaeosurfaces, e.g., Kuk Swamp (Golson et al., 2017).

In some environments, such as the Amazon Basin, certain types of palaeosol have been inferred to be associated with specific types of human activity in the past. Terra preta soils are dark soils enriched with the refuse of former settlements and terra mulata soils are less-enriched versions that represent zones of former cultivation around settlements (Lehmann et al., 2003). Most shifting cultivation practices, however, where plots are cultivated for only 1–3 years prior to abandonment for at least 15 years, may not leave such clear and long-lasting, pedological signatures.

Until the introduction of metal tools within the last century, the tools of plant exploitation and shifting cultivation in the interior of New Guinea were predominantly wooden, such as dibbling and digging sticks. Stone axes would have aided tree felling and ring-barking (Christensen, 1975), and tanged stone blades or “hoes” have been interpreted to represent implements used for digging (Bulmer, 1977). Yet there is no clear archaeological signature for the transition to shifting cultivation in the archaeological record of the interior of New Guinea; although the more widespread adoption of polished/ground stone axe-adzes is associated with increased forest clearance in the Upper Wahgi Valley from c. 7,000–6,500 cal BP (Denham, 2019).

Traditionally, shifting cultivation in New Guinea (as elsewhere) has been inferred from palaeoecological records of pollen and microcharcoal records in cores from lakes and wetlands (Haberle, 2007; Hope and Haberle, 2005). The diagnostic palaeoecological signatures of cultivation in the past are inferred from cumulative and directional changes in the composition of floral assemblages, namely, decreased frequencies of primary forest taxa and concomitant increased frequencies of secondary forest, disturbance, grassland and useful taxa together with increased charcoal particle frequencies (Denham and Haberle, 2008). Basically, palaeoecological records are interpreted in terms of divergence of vegetation communities from that anticipated given the prevailing climate and location; for instance, the extension and maintenance of grasslands during periods of climatic amelioration, including warmer and wetter climates, would suggest human agency in the form of burning, resource intensification or cultivation practices. The absence, though, of primary archaeological evidence can bring into doubt claims for early forms of cultivation, especially shifting cultivation, in the past based on palaeoecological evidence.

Exemplification: re-evaluating early Holocene evidence at Kuk swamp

The wetland archaeological site of Kuk Swamp is located on the floor of the Upper Wahgi Valley, Western Highlands Province, Papua New Guinea. Kuk Swamp is located at c. 1,560 m altitude and forms part of extensive wetlands on the floor of the Upper Wahgi Valley (Figure 5). The wetland was artificially drained in the late 1960s and early 1970s as part of the formation of the Kuk Tea (subsequently Agricultural) Research Station. The archaeological site of Kuk comprises c. 100 ha of the southeastern corner of the former Station and has been reoccupied by the traditional Kawelka owners since the 1990s.

Figure 5

The Kuk landscape comprises wetlands, the Wahgi River and its drainage network, gentle valley slopes and steep valley walls that rise to c. 2,400 m altitude. These landforms sustain diverse wetland, riparian and intra-montane habitats. Today the area is characterised by a slightly seasonal, lower montane humid climate dominated by local orographic effects (Powell et al., 1975). The average annual temperature is 19°C and annual rainfall is c. 2,700 mm (Hughes et al., 1991). A slight dry season occurs between May and June, while soil water content is ordinarily sufficient to maintain plant growth throughout the year (McAlpine et al., 1983).

The multi-disciplinary evidence dating to c. 10,000 years ago comes from the southeastern corner of the wetland margin (Figure 5), comprises palaeochannel and palaeosurface, and is here re-interpreted to investigate whether it plausibly represents shifting cultivation in the past. As with previous interpretations (since Golson, 1977a), the interpretation is not definitive. Even though the multidisciplinary evidence derived from the palaeosurface is consistent with the types of evidence anticipated for shifting cultivation in the past, alternative interpretations are still possible.

The palaeosurface was located on a slightly elevated area of land adjacent to a palaeochannel on the wetland margin (Denham, 2003b, 2004b; Denham et al., 2004, 2017; Golson, 1977a; Golson and Hughes, 1980). The archaeostratigraphic position of palaeochannel and palaeosurface is the same; both overlay black organic clays and are buried by grey clays that commenced deposition at c. 10,000 cal BP (Denham et al., 2003: Tables S2). Six radiocarbon dates on organic materials from the fills of the palaeochannel yielded a combined Bayesian calibrated date of 10,200–9,910 cal BP, shortened to c. 10,000 cal BP here (Denham et al., 2003: Tables S2). Radiocarbon dates from the fills of palaeosurface features fills are more variable (Denham et al., 2003: Table S1), which is anticipated given that they are derived from fine, admixed organics within palaeosols rather than on wood charcoal and bark. Functional associations suggest a period of short-term palaeosurface use while it was locally drier due to effective drainage of this part of the wetland while the adjacent palaeochannel was open (Denham, 2004b; Golson, 1991). Thus, the contemporaneity of palaeochannel and palaeosurface is based on archaeostratigraphic equivalence, comparable antiquity and functional association.

The palaeochannel is almost certainly a natural drainage channel (Denham, 2003b, 2004b; Denham et al., 2004, 2017). Previous interpretations had suggested that it was artificial, namely, it had been dug by people (Golson and Hughes, 1980; see differences of opinion in Denham et al., 2004, 2017). However, there is nothing associated with channel planform, channel gradient or cross-sectional morphology to indicate that the palaeochannel was dug by people (Denham, 2004b).

The fills of the palaeochannel contained admixed soil aggregates, plant remains and woody debris consistent with forest clearance of adjacent land (Denham and Grono, 2017). Soil aggregates are not usually stable in watercourses and are of proximal origin. Furthermore, the abundance and rapid deposition of soil aggregates in the palaeochannel indicate relatively high intensity physical disturbance of adjacent land.

Pollen and microcharcoal records from the fills of the palaeochannel are suggestive of slightly increasing, yet episodic forest disturbance using fire in the vicinity (Figure 6; Haberle et al., 2012, 2017). Paleochannel fills contained phytoliths of bananas and possible taro seeds, suggesting these species were growing locally (Denham, 2008; Lentfer and Denham, 2017; Wilson, 1985). The nature of the human activities suggested by multiproxy analysis of palaeochannel fills are enhanced by detailed archaeological, mixed-method and multiproxy investigation of the adjacent palaeosurface.

Figure 6

Archaeological excavation of an area of higher ground adjacent to the palaeochannel revealed numerous anthropic features comprising a contemporaneous palaeosurface (Figure 7; Denham et al., 2003, 2017; Denham, 2004b). Clusters of features included: intercut pits and stakeholes (Areas D, E) and sinuous microtopographical runnels comprised of intercut series of shallow pits and “basins” (Areas A, B, C; Denham, 2004b; Denham et al., 2004, 2017). The palaeosurface is considered to have been used for only a short-time, perhaps only a few years (Denham et al., 2017, Denham, 2004b, 2018a). After abandonment, the palaeochannel and palaeosurface features were infilled by the grey clay marking the commencement of alluvial fan deposition on the southern catchment margin (Hughes et al., 1991, 2017).

Figure 7

Multiple aspects of the features suggest that they could be associated with plant management, cultivation and harvesting, primarily the morphology, inferred function, nature of fills, palaeoecological evidence, and associated artefacts (following drawn from Denham et al., 2004, 2017). Foremost, the features had clean and defined edges that cut into and contrasted with the underlying stratigraphy, namely, they did not represent shrinkage, compaction or deformation of surface or subsurface sediments. Edge sharpness and smoothness are indicative of dug features rather than tree bowls, root moulds and tree-throw depressions that would all have more irregular edges. The features could have been dug with wooden digging implements, such as digging sticks, or possibly stone hoe-like implements (Bulmer, 2005).

In terms of morphology and function, the stakeholes and shallow pits are consistent with ethnographic accounts of various plant management and cultivation practices. A recurrent feature association of a stakehole with a “basin” (slightly larger pit) may represent the staking and harvesting of a tuberous plant, such as a yam, while other stakeholes are consistent with using stakes to support banana bunches or clumps of cane grasses. Other inter-cut shallow pits and depressions plausibly represent digging, planting and harvesting of tuberous plants and other crops. The runnels were dug to enhance microtopographical drainage within the vicinity, thereby providing growing environments for plants with different edaphic requirements, potentially bananas, cane grasses and yams on slightly drier ground and taro within the wetter runnels, which has been observed at wetlands in the Highlands in the recent past (Denham et al., 2017: Figure 11.11; cf. Golson, 1977a).

The basal fills of the anthropic features contain more admixed sediments than overlying and underlying strata, indicate soil formation and admixture, and are consistent with anthropic disturbance (Denham and Grono, 2017). Multi-scalar stratigraphic investigations of feature fills associated with the palaeosurface indicate that this localised and slightly elevated area was only dry for a short time (Denham et al., 2009; Denham and Grono, 2017). Limited pedogenesis occurred of near-surface sediments to create an immature A/C horizon; a nascent A-horizon formed on near-surface wetland margin sediments, with underlying sediments constituting the C horizon (Figure 8; Denham and Grono, 2017; Mason et al., 2022). X-radiographs and mineral maps clearly show the preservation of soil aggregates, representing a subangular blocky structure (Denham et al., 2009; Mason et al., 2022), and thin section microscopy indicated soil forming processes, some of which are characteristic of cultivation (Denham and Grono, 2017). These pedological characteristics were relatively weakly developed on parent sediments due to subsequent inundation that inhibited further soil formation and aided the preservation of palaeosurface features and palaeosols.

Figure 8

Multiproxy palaeoecological analyses of feature fills complementing those of palaeochannel fills; they indicate burning and transformation of the local environment (Denham et al., 2003, 2009). The diatom assemblages reflect locally drier conditions for a short time, with contemporaneous pollen and microcharcoal showing increased levels of burning and limited transformation of the local vegetation. These multiproxy analyses are securely linked to microstratigraphic investigations of an archaeological feature, thereby enabling an integrated and highly robust basis for interpreting on site vegetation clearance and burning.

A few artefacts were associated from the palaeosurface and palaeochannel. These artefacts comprise three flakes and a pounding/grinding stone. Two artefacts contained starch grain and raphide residues consistent with taro and yam (Figure 6; Fullagar et al., 2006). These artefacts are not directly indicative of cultivation, rather they are more consistent with the processing of locally-grown food prior to consumption. Archaeobotanical investigations of associated fills indicate a range of other food plants were growing locally at this time, including leafy food plants (e.g., Amaranthus sp., Commelina sp., Oenanthe javanica) and edible rushes and grasses (i.e., Typha sp. and cf. Setaria palmifolia, respectively) (Denham et al., 2003). Tools associated with digging for planting and harvesting, such as dibbles and digging sticks, were not identified during excavation of the palaeosurface or palaeochannel. This is not unanticipated because most tools traditionally used in plant exploitation and cultivation are wooden (Golson, 1977b), and these do not ordinarily preserve in archaeological contexts over a few thousand years old at wetland sites in the highlands (Powell, 1982b).

First, archaeobotanical investigations of plant microfossils suggest three key carbohydrate-rich staples of New Guinea agriculture – namely, banana, taro and yam – together with a range of other edible plants were growing in this location at this time (Denham et al., 2003). The co-occurrence of these three carbohydrate rich staples at Kuk is suggestive of deliberate transplantation to this location. It is hard to envisage a scenario in which these three carbohydrate-rich plants would be coincidentally growing together in a short-lived, drier area of the wetland margin otherwise. All three were near the upper limits of their wild altitudinal ranges at this time, where sexual reproduction is inhibited, reinforcing the likelihood of anthropic transplantation by vegetative propagation rather than natural dispersal (Denham et al., 2004).

Second, palaeoecological analyses show anthropic disturbance and periodic burning were not just localised practices occurring in one area of the wetland margin at Kuk, rather they were more general practices that were increasingly disturbing and opening-up the forested floor of the Upper Wahgi Valley. Indeed, similar palaeoecological signals continued for the next 3,000 years to c. 7,000 cal BP, leading to a cumulative reduction in primary forest cover on the valley floor and its gradual replacement with secondary forests, disturbance taxa and grasslands (Haberle et al., 2012). During the same period, c. 10,000 cal BP to c. 7,000 cal BP, an alluvial fan comprised of grey clay was deposited on the wetland margin representing increased rates of erosion within the southern catchment (Hughes et al., 1991). These cumulative and directional palaeoecological and geomorphological signals are interpreted in the Highlands context to be indicative of shifting cultivation, in which the increased density and periodicity of plot cultivation gradually led to degradation of the montane forest and soil erosion (Haberle et al., 2012, 2017).

Although not definitive, the archaeobotanical and palaeoecological evidence enables us to differentiate between plant exploitation and cultivation. First, the multidisciplinary findings are not consistent with adventitious human use of a gap in the montane forest; rather they reflect a greater degree of deliberate human action in terms of the transplantation of key food plants and in terms of the cumulative opening-up and replacement of the primary forest. Gap dynamics would tend to enable regeneration of the forested landscape, rather than its gradual replacement. Second, intentional clearance, burning, digging and staking are plausibly consistent with the management of useful plant resources and gathering activities within a deliberately created patch within the montane forest. Although the cumulative effects of patch creation, maintenance and expansion could potentially account for the observed palaeoecological signals, the transplantation of three significant carbohydrate-rich plants to this location is more characteristic of deliberate cultivation practices.

Based on the foregoing discussion, the most plausible interpretation to account for the observed multidisciplinary evidence is that key food plants were translocated and vegetatively propagated from another location and planted in a prepared plot on the wetland margin at Kuk. The plot had been prepared in terms of the burning and clearing of vegetation and microtopographical manipulation to enable drainage and cultivation of plants with different edaphic requirements. Digging activities and staking were associated with planting, maintenance and harvesting. The plot was used for only a short-time, as is characteristic of shifting cultivation plots in the Highlands today, before abandonment to long fallow.

Living with uncertainty

Although the multidisciplinary evidence associated with the 10,000 year-old “plot” on the wetland margin at Kuk is ambiguous, it plausibly represents an early form of shifting cultivation. Through time, the intensity of shifting cultivation practices in the Upper Wahgi Valley landscape increasingly led to environmental degradation and widespread replacement of forest with grassland by c. 7,000 years ago. The presence of extensive grasslands likely contributed to the initiation of more intensive forms of cultivation using mounds from at least c. 7,000–6,400 cal BP.

The discrimination of shifting cultivation from other plant exploitation practices in the past is always likely to be difficult. There are conceptual and methodological reasons why shifting cultivation may be difficult to distinguish from gap dynamics, patch creation and resource intensification. Direct archaeological evidence and palaeosols associated with former cultivation are unlikely to preserve well, except in a few locations where they are buried and waterlogged, such as wetlands. Multidisciplinary lines of archaeobotanical and palaeoecological evidence can often be interpreted in different ways; they require grounding in archaeological remains associated with former practices to be specific and targeted. Even at Kuk Swamp, where a wide array of detailed multidisciplinary investigations is available, there are still uncertainties whether the 10,000 cal BP palaeosurface represents a resource-rich patch cleared by people in the montane forest, a shifting cultivation plot, or even an alternative type of plant exploitation. Despite these uncertainties, the multidisciplinary method advanced here is designed to stimulate debate about the conceptual, methodological and empirical bases for the identification of shifting cultivation in the past.

References

• 1

  BallardC. (1995). The death of a great land: ritual, history and subsistence revolution in the southern highlands of Papua New Guinea. (unpublished PhD thesis). Australian National University.

  • Google Scholar
  • View reference in article

• 2

  BallardC. (2017). “The wetland field systems of the new Guinea highlands,” in Ten Thousand Years of Cultivation at Kuk Swamp in the Highlands of Papua New Guinea, eds. GolsonJ.DenhamT.P.HughesP.J.SwadlingP.MukeJ. (Canberra: ANU E Press). Terra Australis46, 65–83.

  • Google Scholar
  • View reference in article

• 3

  BallardC.DenhamT. P.HaberleS. G. (2013). “Wetland archaeology in the highlands of new Guinea,” in The Oxford Handbook of Wetland Archaeology, eds. MenottiF.O’SullivanA. (Oxford: Oxford University Press), 231–248.

  • Google Scholar
  • View reference in article

• 4

  BallardC.BrownP.BourkeR. M.HarwoodT. (2005). The Sweet Potato in Oceania: A Reappraisal. Ethnology Monographs 19 and Oceania Monograph 56. Pittsburgh and Sydney: University of Pittsburgh and University of Sydney.

  • Google Scholar
  • View reference in article

• 5

  BarkerG. (2006). The Agricultural Revolution in History. Oxford: Oxford University Press.

  • Google Scholar
  • View reference in article

• 6

  BarrauJ. (1965). Witnesses of the past: notes on some food plants of oceania. Ethnology.4 (3), 282–294. 10.2307/3772988

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  BartonH.DenhamT. P. (2018). Vegecultures and the social-biological transformations of plants and people. Quat. Int.489, 17–25. 10.1016/j.quaint.2016.06.031

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  BartonH.DenhamT. P. (2011). “Prehistoric vegeculture and social life in island Southeast Asia and Melanesia,” in Why Cultivate? Anthropological and Archaeological Approaches to Foraging-Farming Transitions in Southeast Asia, eds. BarkerG.JanowskiM. (Cambridge: McDonald Institute for Archaeological Research). McDonald Institute Monograph, 17–25.

  • Google Scholar
  • View reference in article

• 9

  BourkeR. M. (2001). Intensification of agricultural systems in Papua New Guinea. Asia. Pac. Viewp.42, 219–236. 10.1111/1467-8373.00146

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  BourkeR. M.AllenB. (2009). “Village food production systems,” in Food and Agriculture in Papua New Guinea, eds. BourkeR.M.HarwoodT. (Canberra: ANU E Press), 193–269.

  • Google Scholar
  • View reference in article

• 11

  BourkeR. M.HarwoodT. (2009). Food and Agriculture in Papua New Guinea. Canberra: ANU E Press.

  • Google Scholar
  • View reference in article

• 12

  BrookfieldH. C.HartD. (1971). Melanesia. London: Methuen.

  • Google Scholar
  • View reference in article

• 13

  BulmerS. (1977). “Waisted blades and axes,” in Stone Tools as Cultural Markers: Change, Evolution and Complexity, ed. WrightR.V.S. (Canberra: Australian Institute of Aboriginal Studies), 40–59.

  • Google Scholar
  • View reference in article

• 14

  BulmerS. (2005). “Reflections in stone: axes and the beginnings of agriculture in the central highlands of New Guinea,” in Papuan Pasts: Cultural, Linguistic and Biological Histories of Papuan-Speaking Peoples, eds. PawleyA.AttenboroughR.GolsonJ.HideR. (Canberra: Pacific Linguistics, Research School of Pacific and Asian Studies, The Australian National University). Pacific Linguistics572, 387–450.

  • Google Scholar
  • View reference in article

• 15

  BulmerR. N. H. (1982). “Crop introductions and their consequences in the upper Kaironk valley, Simbai area, Madang province,” in Proceedings of the Second Papua New Guinea Food Crops Conference, Volume 2, eds. BourkeR.M.KesavanV. (Port Moresby: Department of Primary Industry), 282–288.

  • Google Scholar
  • View reference in article

• 16

  CaballeroJ. (2004). “Patterns in human–plant interaction: an evolutionary perspective,” in Paper Presented at the International Society of Ethnobiology, Ninth International Congress, Canterbury, England. Canterbury, 13–17 June 2004.

  • Google Scholar
  • View reference in article

• 17

  ChristensenO. A. (1975). Hunters and horticulturalists: a preliminary report of the 1972–4 excavations in the Manim Valley, Papua New Guinea. Mankind10, 24–36. 10.1111/j.1835-9310.1975.tb00909.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  ClarkeW. C. (1971). Place and People: An Ecology of a New Guinea Community. Berkeley: University of California Press.

  • Google Scholar
  • View reference in article

• 19

  CookC. D. (1999). Pandanus agroforestry of the Amungme in Irian Jaya, Indonesia. Forest Farm Community Tree Res. Rep.4, 95–103.

  • Google Scholar
  • View reference in article

• 20

  DanielsJ.DanielsC. (1993). Sugarcane in prehistory. Archaeol. Ocean.28, 1–7. 10.1002/j.1834-4453.1993.tb00309.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  DenhamT. P. (2003a). Archaeological evidence for mid-holocene agriculture in the interior of Papua New Guinea: a critical review. Archaeol. Ocean.38, 159–176. 10.1002/j.1834-4453.2003.tb00542.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  DenhamT. P. (2003b) The kuk morass: multi-disciplinary investigations of early to mid holocene plant exploitation at Kuk swamp, Wahgi valley, Papua New Guinea. (unpublished PhD thesis). Australian National University.

  • Google Scholar
  • View reference in article

• 23

  DenhamT. P. (2004a). The roots of agriculture and arboriculture in new Guinea: looking beyond austronesian expansion, neolithic packages and Indigenous origins. World. Archaeol.36, 610–620. 10.1080/0043824042000303791

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  DenhamT. P. (2004b). “Early agriculture in the highlands of new Guinea: an assessment of phase 1 at Kuk swamp,” in A Pacific Odyssey: Archaeology and Anthropology in the Western Pacific. Papers in Honour of Jim Specht. Records of the Australian Museum, Supplement 29, eds. AttenbrowV.FullagarR. (Sydney: Australian Museum), 47–57.

  • Google Scholar
  • View reference in article

• 25

  DenhamT. P. (2005). “Agricultural origins and the emergence of rectilinear ditch networks in the highlands of new Guinea,” in Papuan Pasts: Cultural, Linguistic and Biological Histories of Papuan-Speaking Peoples, eds. PawleyA.AttenboroughR.GolsonJ.HideR. (Canberra: RSPAS, ANU). Pacific Linguistics572, 329–361.

  • Google Scholar
  • View reference in article

• 26

  DenhamT. P. (2007). “Early to mid-holocene plant exploitation in new Guinea: towards a contingent interpretation of agriculture,” in Rethinking Agriculture: Archaeological and Ethnoarchaeological Perspectives, eds. DenhamT.P.IriarteJ.VrydaghsL. (Walnut Creek: Left Coast Press), 78–108.

  • Google Scholar
  • View reference in article

• 27

  DenhamT. P. (2008). “Environmental archaeology: interpreting practices-in-the-landscape through geoarchaeology,” in Handbook of Landscape Archaeology, eds. DavidB.ThomasJ. (Walnut Creek: Left Coast Press), 468–481.

  • Google Scholar

• 28

  DenhamT. P. (2009). A practice-centred method for charting the emergence and transformation of agriculture. Curr. Anthropol.50, 661–667. 10.1086/605469

  • CrossRef
  • Google Scholar

• 29

  DenhamT. P. (2011). Early agriculture and plant domestication in new Guinea and island Southeast Asia. Curr. Anthropol.52 (S4), S379–S395. 10.1086/658682

  • CrossRef
  • Google Scholar

• 30

  DenhamT. P. (2015). “The swamp cultivators at Kuk: early agriculture in the highlands of new Guinea,” in The Cambridge World History. Volume II. A World with Agriculture, eds. BarkerG.GoucherC. (Cambridge: Cambridge University Press), 445–471.

  • Google Scholar

• 31

  DenhamT. P. (2016). “Socio-environmental adaptation to the montane rainforests of new Guinea,” in The Routledge Handbook of Bioarchaeology in Southeast Asia and the Pacific, eds. OxenhamM.BuckleyH. (Abingdon, Oxford: Routledge), 409–426.

  • Google Scholar

• 32

  DenhamT. P. (2018a). Tracing Early Agriculture in the Highlands of New Guinea: Plot, Mound and Ditch. Oxford: Routledge.

  • Google Scholar

• 33

  DenhamT. P. (2018b). “Long-term records of human adaptation and tropical rainforest conservation in Papua New Guinea: translating the past into the present,” in Exploring Frameworks for Tropical Forest Conservation, ed. SanzN. (Mexico City: UNESCO), 54–91.

  • Google Scholar

• 34

  DenhamT. P. (2019). “Reconsidering the “neolithic” at Manim Rockshelter, Wurup valley, Papua New Guinea,” in Archaeologies of Island Melanesia: Current Approaches to Landscapes, Exchange and Practice, eds. LeclercM.FlexnerJ. (Canberra: ANU E Press). Terra Australis51, 81–99.

  • Google Scholar

• 35

  DenhamT. P. (2025). Forgetting cane grasses: switching temporal focus to reveal mosaics of Saccharum diversity. Plants People Planet7, 590–602. 10.1002/ppp3.10489

  • CrossRef
  • Google Scholar

• 36

  DenhamT. P.BartonH. (2006). “The emergence of agriculture in new Guinea: continuity from pre-existing foraging practices,” in Behavioral Ecology and the Transition to Agriculture, eds. KennettD.J.WinterhalderB. (California: University of California Press), 237–264.

  • Google Scholar

• 37

  DenhamT. P.GronoE. (2017). Sediments or soils? Multi-scale geoarchaeological investigations of stratigraphy and early cultivation practices at Kuk swamp, highlands of Papua New Guinea. J. Archaeol. Sci.77, 160–171. 10.1016/j.jas.2016.07.007

  • CrossRef
  • Google Scholar

• 38

  DenhamT. P.HaberleS. G. (2008). Agricultural emergence and transformation in the upper wahgi valley during the holocene: theory, method and practice. Holocene18, 499–514. 10.1177/0959683607087936

  • CrossRef
  • Google Scholar

• 39

  DenhamT. P.GolsonJ.HughesP. J. (2004). Reading Early agriculture at kuk (phases 1–3), wahgi valley, Papua New Guinea: the wetland archaeological features. Proc. Prehist. Soc.70, 259–298. 10.1017/S0079497X00001195

  • CrossRef
  • Google Scholar

• 40

  DenhamT. P.GolsonJ.HughesP. J. (2017). “Phase 1: the case for 10,000-year-old agriculture at Kuk,” in Ten Thousand Years of Cultivation at Kuk Swamp in the Highlands of Papua New Guinea, eds. GolsonJ.DenhamT.P.HughesP.J.SwadlingP.MukeJ. (Canberra: ANU E Press). Terra Australis46, 187–200.

  • Google Scholar

• 41

  DenhamT. P.SnidermanK.SaundersK.WinsboroughB.PierretA. (2009). Contiguous multi-proxy analyses (X-radiography, diatom, pollen and microcharcoal) of holocene archaeological features at Kuk swamp, upper Wahgi valley, Papua New Guinea. Geoarchaeology.24, 715–742. 10.1002/gea.20283

  • CrossRef
  • Google Scholar

• 42

  DenhamT. P.HaberleS. G.LentferC.FullagarR.FieldJ.TherinM., et al (2003). Origins of agriculture at Kuk swamp in the highlands of new Guinea. Science301, 189–193. 10.1126/science.1085255

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 43

  DenhamT. P.BartonH.CastilloC.CrowtherA.Dotte-SaroutE.FlorinA., et al (2020). The domestication syndrome in vegetatively propagated field crops. Ann. Bot.125, 581–597. 10.1093/aob/mcz212

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 44

  FlenleyJ. R. (1988). Palynological evidence for land use changes in South-East Asia. J. Biogeogr.15, 185–197. 10.2307/2845059

  • CrossRef
  • Google Scholar

• 45

  FrenchB. R. (2006). Food Plants of Papua New Guinea: A Compendium, revised Edn. Burnie, Tasmania: Food Plants International, Inc.

  • Google Scholar

• 46

  FullagarR.FieldJ.DenhamT. P.LentferC. (2006). Early and mid holocene processing of taro (Colocasia esculenta) and yam (Dioscorea sp.) at Kuk swamp in the highlands of Papua New Guinea. J. Archaeol. Sci.33, 595–614. 10.1016/j.jas.2005.07.020

  • CrossRef
  • Google Scholar

• 47

  FullerD. Q.DenhamT. P.Arroyo-KalinM.LucasL.StevensC.QinL., et al (2014). Convergent evolution and parallelism in plant domestication revealed by an expanding archaeological record. Proc. Natl. Acad. Sci. U. S. A.111, 6147–6152. 10.1073/pnas.1308937110

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 48

  GhazoulJ.SheilD. (2010). Tropical Rain Forest Ecology, Diversity, and Conservation. Oxford: Oxford University Press.

  • Google Scholar

• 49

  GolsonJ. (1977a). “No room at the top: agricultural intensification in the new Guinea highlands,” in Sunda and Sahul: Prehistoric Studies in Southeast Asia, Melanesia and Australia, eds. AllenJ.GolsonJ.JonesR. (London:: Academic Press), 601–638.

  • Google Scholar

• 50

  GolsonJ. (1977b). “Simple tools and complex technology: agriculture and agricultural implements in the new Guinea highlands,” in Stone Tools as Cultural Markers: Change, Evolution and Complexity, ed. WrightR.V.S. (New Jersey: Humanities Press Inc), 154–161.

  • Google Scholar

• 51

  GolsonJ. (1982). “The ipomoean revolution revisited: society and sweet potato in the upper Wahgi valley,” in Inequality in New Guinea Highland Societies, ed. StrathernA. (Cambridge: Cambridge University Press), 109–136.

  • Google Scholar

• 52

  GolsonJ. (1991). “Bulmer phase II: early agriculture in the new Guinea highlands,” in Man and a Half: Essays in Pacific Anthropology and Ethnobiology in Honour of Ralph Bulmer, ed. PawleyA. (Auckland: The Polynesian Society), 484–491.

  • Google Scholar

• 53

  GolsonJ.HughesP. J. (1980). The appearance of plant and animal domestication in New Guinea. J. Soc. Ocean.36, 294–303. 10.3406/jso.1980.3044

  • CrossRef
  • Google Scholar

• 54

  GolsonJ.DenhamT. P.HughesP. J.SwadlingP.MukeJ. (2017). Ten Thousand Years of Cultivation at Kuk Swamp in the Highlands of Papua New Guinea. Terra Australis46. Canberra: ANU E Press.

  • Google Scholar

• 55

  GrivetL.DanielsC.GlaszmanJ. C.D’HontA. (2004). A review of recent molecular genetics evidence for sugarcane evolution and domestication. Ethnobot. Res. Appl.2, 9–17. 10.17348/era.2.0.9-17

  • CrossRef
  • Google Scholar

• 56

  HaberleS. G. (1993). Late quaternary environmental history of the Tari Basin, Papua New Guinea. (unpublished PhD thesis). Australian National University.

  • Google Scholar

• 57

  HaberleS. G. (2007). Prehistoric human impact on rainforest biodiversity in highland New Guinea. Philos. Trans. R. Soc. Lond. B Biol.362, 219–228. 10.1098/rstb.2006.1981

  • CrossRef
  • Google Scholar

• 58

  HaberleS. G.LentferC.DenhamT. P. (2017). “Palaeoecology,” in Ten Thousand Years of Cultivation at Kuk Swamp, eds. GolsonJ.DenhamT.P.HughesP.J.SwadlingP.MukeJ. (Canberra: ANU E Press). Terra Australis46, 145–161.

  • Google Scholar

• 59

  HaberleS. G.LentferC.O’DonnellS.DenhamT. P. (2012). The palaeoenvironments of Kuk swamp from the beginnings of agriculture in the highlands of Papua New Guinea. Quat. Int.249, 129–139. 10.1016/j.quaint.2011.07.048

  • CrossRef
  • Google Scholar

• 60

  HopeG. S.HaberleS. G. (2005). “The history of the human landscapes of new Guinea,” in Papuan Pasts: Cultural, Linguistic and Biological Histories of Papuan-Speaking Peoples, eds. PawleyA.AttenboroughR.GolsonJ.HideR. (Canberra: RSPAS, ANU). Pacific Linguistics572, 541–554.

  • Google Scholar

• 61

  HughesP. J.SullivanM. E.YokD. (1991). Human induced erosion in a highlands catchment in Papua New Guinea: the prehistoric and contemporary records. Z. Geomorphol. Suppl.83, 227–239.

  • Google Scholar

• 62

  HughesP. J.DenhamT. P.GolsonJ. (2017). “Kuk swamp,” in Ten Thousand Years of Cultivation at Kuk Swamp in the Highlands of Papua New Guinea, eds. GolsonJ.DenhamT.P.HughesP.J.SwadlingP.MukeJ. (Canberra: ANU E Press). Terra Australis46, 87–116.

  • Google Scholar

• 63

  KeimA. P. (2012). The pandan flora of Foja-Mamberamo game reserve and Baliem valley, Papua-Indonesia. Reinwardtia13, 271–297.

  • Google Scholar

• 64

  KennedyJ.ClarkeW. C. (2004). “Cultivated landscapes of the southwest Pacific,” in Resource Management in the Asia-Pacific (RMAP) Working Paper 50, (Canberra: RMAP, Australian National University).

  • Google Scholar

• 65

  KistlerL.MontenegroA.SmithB. D.GiffordJ. A.GreenR. E.NewsomL. A., et al (2014). Transoceanic drift and the domestication of African bottle gourds in the Americas. Proc. Natl. Acad. Sci. U. S. A.111, 2937–2941. 10.1073/pnas.1318678111

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 66

  LatinisK. (2000). The development of subsistence system models for island Southeast Asia and near oceania: the nature and role of arboriculture and arboreal-based economies. World Archaeol.32, 41–67. 10.1080/004382400409880

  • CrossRef
  • Google Scholar

• 67

  LebotV.PranaM. S.KreikeN.van HeckH.PardalesJ.OkpulT., et al (2004). Characterisation of taro [Colocasia esculenta (L.) Schott] genetic resources in Southeast Asia and oceania. Genet. Resour. Crop Evol.51, 381–392. 10.1023/B:GRES.0000023453.30948.4d

  • CrossRef
  • Google Scholar

• 68

  LehmannJ.KernD.GlaserB.WoodsW. (2003). Amazonian Dark Earths: Origins, Properties and Management. Dordrecht: Kluwer Press.

  • Google Scholar

• 69

  LentferC.DenhamT. P. (2017). “The archaeobotany of kuk,” in Ten Thousand Years of Cultivation at Kuk Swamp in the Highlands of Papua New Guinea, eds. GolsonJ.DenhamT.P.HughesP.J.SwadlingP.MukeJ. (Canberra: ANU E Press). Terra Australis46, 163–184.

  • Google Scholar

• 70

  MalapaR.ArnauG.NoyerJ. L.LebotV. (2005). Genetic diversity of the greater yam (Dioscorea alata L.) and relatedness to D. Nummularia lam. and D. Transversa br. as revealed with AFLP markers. Genet. Resour. Crop Evol.52, 919–929. 10.1007/s10722-003-6122-5

  • CrossRef
  • Google Scholar

• 71

  MartinG.CottinA.BaurensF.-C.LabadieK.HervouetC.SalmonF., et al (2023). Interspecific introgression patterns reveal the origins of worldwide cultivated bananas in New Guinea. Plant J.113, 802–818. 10.1111/tpj.16086

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 72

  MasonJ.LinE.GronoE.DenhamT. P. (2022). QEMSCAN® analysis of clay-rich stratigraphy associated with early agricultural contexts at Kuk swamp, Papua New Guinea. J. Archaeol. Sci. Rep.42, 103356. 10.1016/j.jasrep.2022.103356

  • CrossRef
  • Google Scholar

• 73

  McAlpineJ. R.KeigG.FallsR. (1983). Climate of Papua New Guinea. Canberra: CSIRO and ANU Press.

  • Google Scholar

• 74

  MukeJ.DenhamT. P.GenorupaV. (2007). Nominating and managing a world heritage site in the highlands of Papua New Guinea. World Archaeol.39, 324–338. 10.1080/00438240701464947

  • CrossRef
  • Google Scholar

• 75

  PerrierX.De LangheE.DonohueM.LentferC.VrydaghsL.BakryF., et al (2011). Multidisciplinary perspectives on banana (Musa spp.) domestication. Proc. Natl. Acad. Sci. U. S. A.108, 11311–11318. 10.1073/pnas.1102001108

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 76

  PowellJ. M. (1970). The impact of man on the vegetation of the mount hagen region, new Guinea. (unpublished PhD thesis). Australian National University.

  • Google Scholar

• 77

  PowellJ. M. (1982a). “The history of plant use and man’s impact on the vegetation,” in Biogeography and Ecology of New Guinea. Volume 1, ed. GressittJ. L. (The Hague: Junk), 207–227.

  • Google Scholar

• 78

  PowellJ. M. (1982b). Plant resources and palaeobotanical evidence for plant use in the Papua New Guinea highlands. Archaeol. Ocean.17, 28–37. 10.1002/j.1834-4453.1982.tb00036.x

  • CrossRef
  • Google Scholar

• 79

  PowellJ. M.KulungaA.MogeR.PonoC.ZimikeF.GolsonJ. (1975). Agricultural Traditions in the Mount Hagen Area. Port Moresby: Department of Geography, UPNG (Occasional Paper No. 12).

  • Google Scholar

• 80

  RappaportR. (1968). Pigs for the Ancestors: Ritual in the Ecology of a New Guinea People. New Haven: Yale University Press.

  • Google Scholar

• 81

  SillitoeP. (1983). Roots of the Earth. Kensington: UNSW Press.

  • Google Scholar

• 82

  SnidermanJ. M. K.FinnJ.DenhamT. P. (2009). A late holocene palaeoecological record from Ambra crater in the highlands of Papua New Guinea and implications for agricultural history. Holocene19, 449–458. 10.1177/0959683608101394

  • CrossRef
  • Google Scholar

• 83

  StoneB. C. (1982). “New Guinea Pandanaceae: first approach to ecology and biogeography,” in Biogeography and Ecology of New Guinea, ed. GressittJ.L. (The Hague: W. Junk), 401–436.

  • Google Scholar

• 84

  WilsonS. M. (1985). Phytolith evidence from Kuk, an early agricultural site in New Guinea. Archaeol. Ocean.20, 90–97. 10.1002/j.1834-4453.1985.tb00115.x

  • CrossRef
  • Google Scholar

• 85

  YenD. E. (1973). The origins of Oceanic agriculture. Archaeol. Phys. Anthropol. Ocean.8, 68–85.

  • Google Scholar

• 86

  YenD. E. (1985). “Wild plants and domestication in Pacific islands,” in Recent Advances in Indo-Pacific Prehistory, eds. MisraV.N.BellwoodP. (New Dehli: Oxford and IBH Publishing), 315–326.

  • Google Scholar