Dating the Neolithic of South India: new radiometric evidence for key economic, social and ritual transformations.

Fuller, Dorian Q. ; Boivin, Nicole ; Korisettar, Ravi 等

The Neolithic period in South India is known for its ashmounds, superseded (in its Iron Age) by megalith builders with craft specialisation. Thanks to a major radiocarbon dating programme and Bayesian analysis of the dates, the authors have placed this sequence in a new chronological framework: the ashmounds, formed by burning cattle dung, are created by a few generations of people. In many cases the mounds are then succeeded by villages, for which they may have acted as founding rituals. The new tightly dated sequence also chronicles the cultivation of particular crops, some indigenous and some introduced from Africa.

Keywords: Neolithic, South Asia, South India, ashmounds, megaliths

Introduction

The Deccan plateau of South India is a large, arid region featuring rich Neolithic period remains (see Figure 1). Focused in particular on the often spectacular granite hills that dot an otherwise largely featureless landscape, South Indian Neolithic sites reveal a unique manifestation of the transition to sedentism together with early pastoral mobility. The region's ashmound sites (Figure 2), consisting of large accumulations of vitrified and non-vitrified cattle dung ash are argued to have been created during the course of ritual activities (Allchin 1963; Paddayya 1991-92; 2000-2001; Korisettar et al. 2001a; Boivin 2004a; Johansen 2004). South India is also of interest for its subsequent Megalithic phase, which is marked by the creation of a large number and diverse variety of stone built burial monuments (Figure 3) and stone alignments (Allchin 1956; Leshnik 1974; Sundara 1975). These are generally thought to testify to a more complex and hierarchical society (Moorti 1994; Brubaker 2001; Mohanty & Selvakumar 2001), and are attributed by some to the arrival of immigrants into the area (e.g. Leshnik 1974). However, the relationship between the period of Megalithic burials, focused on the first millennium BC, and the Neolithic period, which appears to fall within the third and second millennia BC, remains unclear. In addition, phasing within these periods, especially the two millennia of the Neolithic period, and its implications for changing social and economic systems, is still poorly resolved.

[FIGURE 1 OMITTED]

After a lull in Southern Neolithic studies from the early 1970s until the early 1990s, the last decade has seen a re-emergence of interest in the Neolithic developments of South India. Recent investigations have focused, for example, on elucidating agricultural developments and origins (Fuller et al. 2001; 2004; Fuller 2003a; Fuller & Korisettar 2004), lithic production techniques (Paddayya 1993a; 1993b; DuFresne et al. 1998; Brumm et al. 2006), the relationship between sites and landscapes (Boivin 2004a), the role of rock art and ringing rocks (Boivin 2004b), and the early distribution of Dravidian languages (Fuller 2003b; Southworth 2005). Interestingly, recent years have also seen the emergence of a number of debates, focused on the nature of Neolithic site occupations, the evidence for different site types, and potential models of how these sites fit together into a settlement system. In particular, the ashmound debate has polarised those who argue that ashmound sites are always seasonal encampments of mobile herders, or transhumant segments of agricultural villages (e.g. Allchin 1963; Fuller et al. 2001; Korisettar et al. 2001 a), and those who regard ashmounds as a component of typical sedentary village sites (e.g. Paddayya 1991-92; 2000-2001; 2002; Devaraj et al. 1995; Johansen 2004).

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

Much of the recent debate in Southern Neolithic archaeology relies on evidence collected during excavations and research conducted from the late 1950s to the early 1970s (see Korisettar et al. 2001 a for a review). It was during this period that radiocarbon dating techniques were first applied in Indian archaeology. Since this early period, however, Southern Neolithic chronology has been little modified or refined, although it is increasingly clear that numerous key issues require refined chronological understanding. The sparse dating evidence, and tacit acceptance that wide error bars indicate long, continuous phases, has tended to emphasise site longevity and continuity. Recent discussions of ashmounds as Neolithic monuments (e.g. Boivin 2004a; Johansen 2004), for example, are based on an understanding that ashmound sites were in use over long time periods. In addition, the settlement-subsistence model of Fuller et al. (2001) assumes contemporaneity of several hilltop village sites and ashmounds. These models, however, are not clearly supported by all available dating evidence, and the need to establish the contemporaneity of different sites objectively and the length of the formation of sites, including ashmounds, is acute. In addition trends in settlement pattern through the course of the Neolithic require critical assessment.

The present paper offers a new chronological model for the Southern Neolithic that helps to resolve many of these outstanding problems. It not only adds 57 new AMS dates to the existing corpus of 59 published radiocarbon dates (Figure 4), but also offers an assessment of all of the dates based on the application of Bayesian statistical models. The resulting chronologies demand a re-evaluation of accepted ideas about the timing and significance of the distinctive ashmound sites of South India, as well as a re-interpretation of their role in marking the social landscape of prehistoric South India. It also provides direct dates for the use of a number of important crops, both indigenous and introduced, and clarifies the period of the transition to the Megalithic period, associated with the emergence of craft specialisation and social hierarchy.

Old dates, new dates and new calibrations

Most previously-published radiocarbon dates for the Southern Neolithic represent conventional bulk charcoal dates, many of which were obtained in the 1960s and 1970s. They are distributed across 17 sites, half of which have only one or two dates (see Figure 4). The earlier dates, together with ceramic typology and stratigraphic evidence (derived mainly from excavations at Umur and Piklihal (see Allchin 1963)), formed the basis for the construction of the conventional four period chronology for this region: three periods of the Neolithic followed by a transitional Neolithic to Megalithic phase (Allchin & Allchin 1968; 1982).

[FIGURE 4 OMITTED]

Our new dates derive from excavations and section cleanings carried out as part of several ongoing projects in South India, but especially the Sanganakallu-Kupgal Project (Fuller et al. 2001; Korisettar et al. 2001b; Boivin 2004a, 2004b; Boivin et al. 2005; Brumm et al. 2006), involving work at a number of sites in and around the cluster of hills bracketed by the villages of Sanganakallu and Kupgal near the town of Bellary in eastern Karnataka (Figure 1). Of particular interest in the context of the present discussion are the sites on Sannarachamma and Hiregudda, two hills that form part of a larger cluster in this area. The project has provided 8 new dates for the repeatedly-studied occupation at Sannarachamma, and 13 new dates for Hiregudda, a site that has not previously been dated. Other dates have been acquired within the context of a project aimed at elucidating the origins of agriculture in South India (for details see Fuller et al. 2001; 2004; Fuller 2003b; Korisettar 2004). These include new dates for Hallur, Tekkalakota and for several previously undated sites (Figure 4; see also Tables 3-9 at http://www.antiquity.ac.uk/projgall/fuller).

It has recently become increasingly clear that radiocarbon data can be refined by combining information about other dates and about archaeological context. This Bayesian analysis is facilitated by OxCal software (Bronk Ramsey 1995; 2001; 2003), and numerous case studies over more than a decade have now demonstrated their utility in achieving a better understanding of the significance of radiocarbon dates (Buck et al. 1991; 1992; Bayliss et al. 1997; Zeidler et al. 1998; Bronk Ramsey 2000). The present paper will show the importance of Bayesian statistics for refining our understanding of South Indian Neolithic chronology, and revising models of economic, social and ritual change in late South Indian prehistory.

The Bayesian approach allows two sets of information to be combined: the radiocarbon dates, and models of sequence and phase derived from archaeological observations. Dates from the same stratum, or in some cases artefactual periods, are grouped into phases, while sequence ordering is provided by stratigraphy. By including 'boundaries' in the model between key phases in a given site's history it has been possible to model the probable period of key archaeological transitions. OxCal indicates (as a percentage) an agreement index between the 'prior' information, consisting of the individual calibrated dates, and the Bayesian model (or 'posterior probability'). As will be seen, especially for the new dating evidence, the models below have high indices of agreement. In addition, the fact that similar conclusions regarding the age of key regional period transitions is supported by models from multiple sites suggests that our results are robust. All of the new radiocarbon data are reported in full in supplemental online tables (http://www.antiquity.ac.uk/projgall/fuller), while published dates are incorporated in some calibration figures, with sources cited in the captions.

Dating the beginnings of the ashmound tradition and establishing general chronology for the Southern Neolithic

The beginnings of the Southern Neolithic and of the ashmound tradition in particular are often dated by reference to the site of Kodekal, where one of the earliest layers produced charcoal dating to c. 3000-2800 BC (Paddayya 1973: 64). Similarly, the earliest ceramic layers at Watgal date from 2900-2800 BC (Devaraj et al. 1995). Unfortunately at neither of these sites is ashmound formation dated, although they do at least provide the beginning point for a chronology of the Neolithic (ceramic) cultural sequence in the region. While Kodekal is indeed an ashmound site, the dated charcoal actually comes from a lower deposit that predates the ashmound. At both these sites, the material from the earliest phases suggests intermittent, perhaps seasonal use of these sites. At Watgal, more substantial and probably sedentary occupation developed later, beginning c. 2200 BC.

The earliest date for formation of an ashmound comes from Utnur. Here, excavation established the presence of cattle hoof prints, a sequence of posthole defined pens, and evidence of episodes of dung burning within these pens, leading to the build-up of an ashmound (Allchin 1963). Three radiocarbon dates are available from Utnur, but how one interprets these is a matter of the assumptions that one brings to the calibration. Do we assume a long span of occupation or a briefer period? When the calibration probabilities are simply added, there is a focus on the centuries between 2800 and 2200 BC, which is very much how most chronologies of South India have treated the site, thus implying a 600-year long ashmound site.

When the dates and stratigraphy are considered together, however, a much shorter timespan can be suggested for Utnur (Figure 5). While the earliest date suggests the site was founded in the first half of the third millennium BC, conversion to an ashmound (site Phase IC)--indicated unequivocally with a levelling, digging of postholes for a pen, and dung burning--occurs later, c. 2600-2500 BC. After several phases of ashmound formation the site is abandoned, by c. 2200 BC. Such a model assumes two phases of ashmound creation over a period of perhaps 300 years. This is in line with the generally held assumption that ashmounds represent long-term, cyclically-used cattle-camps and ritual sites, in a Neolithic landscape of seasonal transhumance and cultivation (Allchin 1963; Fuller et al. 2001; Korisettar et al. 2001a; Boivin 2004a; Johansen 2004). However, the dates suggest a much shorter life span for the site than the 600 years normally quoted. It is worth emphasising that the modelled sequence could be suggested to be even shorter, if we take the boundaries to be 2500, 2400 and 2300 BC, for example. This example highlights the need for chronometric evidence to refine southern Neolithic chronology.

Beyond the ashmound tradition core region, other regional manifestations of the South Indian Neolithic appear to begin later (Figure 6). In southern Karnataka and adjacent Tamil Nadu, available evidence indicates the founding of sites only during Period II, between 2200 and 1800 BC. A later extension of the Neolithic into the Kunderu river valley, known for its distinctive painted pottery (Patupadu Ware), begins from 1900 BC, based on dates from Sanyasula cave, with village sites in the Kunderu plain from 1700 to 1500 BC. On current evidence, most Neolithic sites of this regional tradition appear to have been abandoned by c. 1400 BC, which associates Neolithic abandonment in the Kunderu river with the period of major social transformation further west, the Neolithic to Megalithic transition.

Budihal and the question of ashmound-village contemporaneity

To date, the most thoroughly excavated ashmound site is Budihal-south (there are three other ashmounds at the Budihal site), where recent excavations have also provided a large number of radiocarbon dates (Paddayya 1993a; 1993b; 1998; 2000-2001; 2002). Budihal is central to a new hypothesis of Southern Neolithic settlement that sees ashmound sites not as a distinct category of seasonal pastoral camp, but rather as part of sedentary village occupations inhabited all the year round (Paddayya 1991-92; 2000-2001). This hypothesis has been criticised on the basis of field observations at a number of other ashmounds sites that lack any substantial non-ashmound occupation deposits (e.g. Kudatini, Utnur, Godekal), in stark contrast to the deeply stratified occupations that are a feature of many hilltop sites (e.g. Sannarachamma, Hiregudda, Tekkalakota, Hatibellagallu) (Fuller et al. 2001a; Korisettar et al. 2001a). But the idea that ashmounds are typical features of Southern Neolithic sedentary villages has been accepted by several authors (Deveraj et al. 1995; Johansen 2004).

The notion that all ashmounds were sites of year-round occupation during the Neolithic originally emerged when it was found that the cattle-pen and ashmound formation at Budihal was contemporaneous with an occupation area that included evidence of round houses. Our re-analysis of the dates from Budihal radically challenges this interpretation.

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

When dates from the ashmound are simply calibrated and summed, as are those from the village settlement, they indeed show overlapping distributions, of c. 2450-2100 BC for the ashmound, and c. 2450-1600 BC for the village area. Use of the dates in this way, however, fails to take into account the available archaeological information, such as the existence of multiple dates and the stratigraphic evidence. Re-analysis using Bayesian statistics indicates a substantial accumulation of dung and the creation of an ashmound over a quite brief period (less than a century; 3-4 human generations), between 2300-2200 BC (Figure 7). While the village occupation area may also begin to be occupied during this period, the village develops and continues until 1700 BC, in the shadow of an ashmound that has long ceased to accumulate. Of additional interest is the apparent temporal association of the ashmound accumulation and a nearby animal butchery floor (Paddayya et al. 1995; Paddayya 1998:150-51), which might relate to feasting in relation to ashmound formation. This would seem to constitute further evidence to support the interpretation ofashmounds as special-function (ritual) seasonal gathering sites rather than regular year-round habitation locales. This evidence also suggests a site life history that moved from pastoral penning and ritual burning, i.e. ashmound formation, to one of a village that is associated with the fixed 'monument' of an ashmound.

Sannarachamma hill: the emergence of a pattern

Recent excavations at Sannarachamma hill near the modern-day village of Sanganakallu provide an opportunity to date ashmound deposits that are sealed within a stratigraphic sequence, as well as the transition to the Megalithic period (Figure 8). This site received much earlier attention as a representative hilltop village of the Southern Neolithic (Subbarao 1948; Ansari & Nagaraja Rao 1969), and has also served as a key site in recent archaeobotanical studies (Fuller eta/. 2001; 2004). More recent excavations at Sannarachamma are providing a representative assemblage of lithic, ceramic, bone and plant evidence from complete sieving and large-scale flotation (Boivin eta/. 2005). Of particular interest here are the insights that these renewed investigations are providing into the changing nature of occupation and deposit formation at the sire. Especially relevant has been the elucidation of a thick ashmound layer sealed by later Neolithic occupation deposits. It is now clear that some of the earliest Neolithic activities at Sannarachamma involved the creation of an ashmound in the centre of the hilltop plateau. This ashmound was subsequently concealed as later occupation deposits covered it.

When the new stratigraphic evidence and dates from Sannarachamma are taken into consideration, the ashmound (both initial and later) can be seen to represent a fairly short phase of activity early in the site's life history, between 1950 and 1750 BC (Figure 9). The previously reported radiocarbon dates (Ansari & Nagaraja Rao 1969) for the site fit with this general model for the Sannarachamma ashmound. This means that the entire period of ash formation covers perhaps 200 years or even less, suggesting once again a fairly short period, of up to 8-10 human generations, for the formation of a substantial ashmound. Of particular significance is the implication that after ashmound formation ceased, intensive village occupation developed at the site for many centuries (perhaps 500-700 years), and subsequently obscured the evidence for the ashmound. Thus, as at Budihal, the ashmound can be suggested to represent an initial stage in the formation of the Neolithic settlement, raising questions about the symbolic role the ashmound may have played in making or marking (perhaps 'purifying') a place for long-term human habitation.

[FIGURE 8 OMITTED]

This emerging model indicates that both sides in the ashmound-village debate require some revision. The 'ashmound' and 'village' actually seem to represent distinct occupation phases, each very different in nature. This is consistent with evidence gathered from other sites by Korisettar and colleagues (2001a) in support of Allchin's (1963) original inference that ashmounds represent seasonal encampments of people and their herds. Nonetheless, it is now also clear that ashmounds do subsequently become villages. However, these villages replace, rather than coexist with, ashmound activities. In the case of Sannarachamma, where occupation on the hilltop plateau was more confined, habitation deposits developed on top of the ashmound, while at Budihal they extended laterally across the plain. Several other recently explored sites also show ashmound and sedentary village phases in stratigraphic sequence (see below). This suggests that there was a widespread pattern of ashmound creation followed by sedentary village occupation during the Neolithic in the southern Deccan. Thus ashmound creation can be seen to be a recurrent phase in Neolithic site creation and life-history in this region.

Short duration versus recurrent ashmounds

The data discussed thus far imply that ashmound creation activities were relatively short-lived on particular sites. That is, the ritual activities that are argued to have led to such deposits (Allchin 1963; Boivin 2004a), and that involved the cyclical or episodic burning of accumulated, or perhaps heaped, cattle dung, seem to have taken place over a restricted number of human generations. Subsequent to this, these ashmound sites were either abandoned--as at Utnur--or became different kinds of sites, as sedentary village occupation developed--as at Sannarachamma and Budihal, a which point ashmound rituals may have moved to other localities.

This may not, however, be the story for all ashmounds. Some particularly massive ashmounds, like those at Kudatini and Palavoy, contain layers that may represent natural soil formations, and could indicate extended phases of abandonment during which natural pedogenesis took place. Allchin also suggested on the basis of surface ceramics that the Kudatini ashmound appears to have spanned several periods of the Neolithic (perhaps from Period I to III; see Allchin 1963). Dates from the few layered ashmounds for which radiocarbon evidence exists are shown in Figure 10. While those from Terdal can be interpreted in terms of a sequence of only two centuries or so, the dates from Palavoy strongly suggest a much longer site span. Overall then it is clear that some ashmounds were indeed formed over quite long time periods, although the possible existence of abandonment layers may indicate that any particular episode of ashmound formation was actually relatively limited. What remains to be resolved is the question of why ashmound formation was re-initiated at some ashmound sites like Palavoy and Kudatini, while other locales were abandoned or became village settlements after one or two centuries of dung-burning activity.

Hiregudda hill: from ashmound to stone axe workshop

Along with Sannarachamma, another major Neolithic site in the Sanganakallu-Kupgal area is Hiregudda (or Kupgal Hill). A medium-sized hilltop plateau referred to as Area A appears to have been the most intensively occupied locale on Hiregudda, and contains the hill's deepest stratigraphy, accompanied by evidence of an ashmound and stone tool production centre. Neolithic activities also took place on other areas of the hill, however, and rock art sites, lithic production areas, stone-walled features and habitation deposits are distributed across most of its slopes and plateaux. Most of the dates come from three adjacent but inter-related stratigraphic sequences in Area A (Figure 11). The chronological model for Hiregudda Area A (Figure 12) suggests a period for the main occupation from 1700 to 1500 BC, with the earlier ashmound represented by some redeposited ash that gives a pre-1700 BC date. This sequence is therefore similar to that on the nearby Sannarachamma hilltop. After 1500 BC, there may have been a hiatus in occupation of a century or so, although further dating evidence is needed to confirm this.

The subsequent transition to the Megalithic period at Hiregudda is associated with increasingly specialised craft production. The large-scale production of stone axes (see Brumm et al. 2006) is correlated both stratigraphically and chronometrically with the production of Megalithic-type wheel-finished pottery (black and red ware), dated to 1400-1300 BC. It is also to this period that several child urn-burials from Hiregudda Area D likely date. Such urn-burials are common during the Megalithic period (Moorti 1994; Brubaker 2001), but are also well-known from several Southern Neolithic contexts, and represent an element of cultural continuity between the two phases Korisettar et al. 2001a).

[FIGURE 9 OMITTED]

[FIGURE 10 OMITTED]

[FIGURE 11 OMITTED]

[FIGURE 12 OMITTED]

Taken together, the evidence from Hiregudda and Sannarachamma indicate that the Neolithic period in the Sanganakallu-Kupgal area begins around 1950 BC with ashmound creation activities at Sannarachamma and Hiregudda that ceased by c. 1750 BC. The end of the Neolithic occurs sometime around the thirteenth century BC with an intensification of craft production activities that is also associated with a new phase of megalith-building, and ultimately the abandonment of hilltop village sites.

The Neolithic-Megalithic transition

As the above discussion suggests, the new radiometric evidence discussed here is important for addressing the relationship between the Neolithic and Megalithic periods. While the Megalithic period is considered by some to be synonymous with the early Iron Age, its chronology is actually still poorly resolved. Relatively few megaliths have been directly dated (Brubaker 2001). Dates from individual graves are in any case not necessarily helpful for identifying the start of this phase, since it is generally accepted that megalithic burial traditions continued until the end of the first millennium BC and even into the first centuries AD (Leshnik 1974; Moorti 1974; McIntosh 1985; Brubaker 2001; Mohanty & Selvakumar 2001). Nevertheless, some available early dates from graves suggest that in northern Karnataka, megalithic burial practices had already begun by 1400-1300 BC. Four thermoluminescence dates on ceramics from burials are available from the site of Kumaranahalli in northern Karnataka (Singhvi et al. 1991; Brubaker 2001: 294-95), and focus on 1400-1300 BC.

The later part of the sequence at Sannarachamma also preserves evidence of the transition to the Megalithic period (SGK 98A-4, context 1157), and also places it between 1400 and 1250 BC (Figure 9). It is significant, however, that these layers have produced no evidence for iron objects, and thus the beginnings of the Megalithic period, defined in ceramic terms, are not necessarily equivalent to the beginnings of iron use.

That the earliest Megalithic relates to the end of ashmound tradition is confirmed by evidence from the site of Velpumudugu. Here the earliest layer is composed of ashmound material, sealed by subsequent occupation, and finds include megalithic black and red ware. Two new AMS dates confirm that this ashmound was formed and sealed during this same transitional period of 1400-1300 BC.

At the site of Hallur, located on the upper Tungabhadra River, continuity of occupation was found through the Iron Age (Nagaraja Rao 1971). The radiocarbon evidence provides a clear framework for this, with occupation from c. 2000 BC, but clearly through several phases between 2000 and 1000 BC (Figure 13). The calibration model suggests that the transition to site Phase 3, which included both wheelmade Black-and-Red ware and a few finds of iron, focuses on 1200-1100 BC. This might suggest that in some regions, the transition to the Megalithic period is somewhat later and does in fact correlate with the spread of iron technology.

[FIGURE 13 OMITTED]

Discussion: implications for economic, social and ritual transformations

The new AMS radiocarbon dates and the reassessment of Southern Neolithic chronology (Table 1) demand a re-evaluation of the significance and role of the ashmounds in South Indian Neolithic society (Table 2). While their status as ritual formations is generally accepted, most models had previously assumed gradual accumulation over an extended period (e.g. Allchin 1963; Paddayya 1991-92; Korisettar et al. 2001a; Boivin 2004a; Johanson 2004). The available evidence now suggests that many, if not most, ashmounds were formed over a fairly short period of time, perhaps a few human generations, and are thus the outcome of much more intensive activities. It also appears that the formation of ashmounds was not, as previously thought, restricted to a particular period within the Southern Neolithic, but was rather a locally-contingent element in the life history of individual sites. The creation of many ashmounds through repetitive, symbolic dung-burning events thus takes on significance as an element of local 'performances' that set the stage for the establishment of village sites. Still, some ashmound sites never developed into villages, and some may have been formed over a more extended period of time, through several alternating phases of ash formation and abandonment. What therefore remains enigmatic is the significance of these differences in site life histories: why some ashmounds, such as Utnur, were abandoned, while others, like Budihal, Hiregudda and Sannarachama, were transformed into villages. The available chronometric evidence highlights the importance of assessing individual ashmounds within the particular social and economic context of individual sites and site groups, but also attests to a distinctive and long-lasting tradition in the creation of settlement spaces and places during the Neolithic of South India.

The new dates also hold implications for our understanding of crop cultivation and diffusion, and hence patterns of trade and interaction. Direct dates on the identified seeds of several crop species (see Table 1) provide the first direct evidence for the antiquity of their cultivation in South India. Beyond those indigenous domesticates, we also have the first direct dates on introduced species. For wheat and barley, early domesticates of Southwest Asia that were staple crops of the Indus civilisation (Zohary & Hopf 2000; Fuller & Madella 2001), we have direct dates back to 1900-1800 BC at Sannarachamma, 1800- 1700 BC at Piklihal, c. 1700 BC at Hiregudda. The adoption of these cereals correlates with the adoption of new ceramic jar forms, and has been argued to represent a new development in cuisine, or beverage consumption, that played a role in creating a new arena of social differentiation in South India (Fuller 2005). In addition we have the first direct AMS dates on a crop of African origin in India, with several dates on hyacinth bean (Lablab purpureus), between 1600 and 1400 BC. At Hallur, the stratigraphically lowest sample with Lablab also contained Pearl millet (Pennisetum glaucum) a crop of West African origin (Fuller 2003c; Fuller et al. 2004). These data suggest the growing geographic exchange network of which South India had become a part. By the end of the Neolithic period textile production had also begun, indicated by spindle whorls (Ashmound Phase III) and our direct date on cotton from Hallur (900-800 BC). This suggests the early beginnings of the South Indian cotton textile industry, which came to be a major export of peninsular India by the historical (Roman) period (Casson 1989).

The adoption of new crops, including a non-subsistence crop like cotton, highlights the development of long-distance exchange networks during the course of the Neolithic. The elaboration of these networks is clearly associated with the development of specialised craft production activities, evidenced not only by the introduction of new ceramic technologies and production regimes, but also, at Hiregudda in particular, by intensive, standardised axe production activities in the fourteenth century BC. It is now clear that the transformation from the Neolithic to the Megalithic period needs to be seen in terms of internal economic and social transformations rather than as the product of the arrival of new groups into the region. What remains to be more adequately worked out is how Neolithic cultural practices, which involved creating places for settlement through dung-burning, ashmound- forming rites, came to be abandoned, while labour and ritual practice increasingly focused on the burials of a small segment of society in megaliths, during the same period that specialised craft production increased. The chronological evidence suggests that the origins of this political economy must be sought in the transformations of the late Neolithic. Further study of ashmounds and related sites is therefore of significance for understanding not only the Neolithisation of South India, in terms of the establishment of agricultural settlements and sedentism, but also the subsequent creation of political economies, featuring craft specialisation, trade and social hierarchy.

Acknowledgements

We thank the archaeological survey of India, and their Director General C. Babu Rajiv, for permission to obtain AMS dates. Funding for fieldwork and associated analyses has been provided by the British Academy, Arts and Humanities Research Council, McDonald Institute for Archaeological Research, Leverhulme Trust, and Karnatak University. Archaeobotanical research has been funded by the Leverhulme Trust. We gratefully acknowledge also the hard work of those who have participated in the Sanganakallu-Kupgal project and other projects on which this paper draws (including the Kurnool District Project), including Eleni Asouti, Janardhana B., Adam Brumm, Subhas Chincholi, Franca Cole, Nick Drake, Neha Gupta, Anitha H.M., Emma Harvey, Deepak Havanur, Linganna K., Udayashankar K., Jinu Koshy, Helen Lewis, Lindsay Lloyd-Smith, Kalyan Malagyannavar, Paul Masser, Stephanie Meece, Justin Morris, Michael Petraglia, Pragnya Prasanna, Shankar Pujar, Arun Raj, Ramadas, Nick Rees, Dave Robinson, EC. Venkatasubbaiah, S.R. Walimbe, and Paula Whittaker.

References

AGRAWAL, D.P. 2002. Dating of Indian Stone Age and Quaternary deposits, in S. Settar & R. Korisettar (ed.) Indian Archaeology in Retrospect, volume III. Archaeology and Interactive Disciplines: 467-84. New Delhi: Manohar.

ALLCHIN, F.R. 1956. The Stone Alignments of Southern Hyderabad. Man 56:133-56.

--1963. Neolithic Cattle Keepers of South India: a Study of the Deccan Ashmounds. Cambridge: Cambridge University Press.

ALLCHIN, B. & R. ALLCHIN. 1968. The Birth of Indian Civilization. Harmondsworth: Penguin.

--1982. The Rise of Civilization in India and Pakistan. Cambridge: Cambridge University Press (also published in 1983 in New Delhi: Select Book Service Syndicate).

ANSARI, Z.D. & M.S. NAGARAJA RAO. 1969. Excavations at Sanganakallu--1964-65. Poona: Deccan College.

BAYLISS, A., C. BRONK RAMSEY & F.G. McCORMAC. 1997. Dating Stonehenge, in B. Cunliffe & C. Renfrew (ed.) Science and Stonehenge (Proceedings of the British Academy 92): 39-59.

BOIVIN, N. 2004a. Landscape and cosmology in the south Indian Neolithic: New perspectives on the Deccan ashmounds. Cambridge Archaeological Journal 14(02): 235-57.

--2004b. Rock art and rock music: petroglyphs of the South Indian Neolithic. Antiquity 78: 38-53.

BOIVIN, N., R. KORISETTAR & D.Q. FULLER. 2005. Further research on the Southern Neolithic and the Ashmound Tradition: The Sanganakallu-Kupgal Archaeological Research Project, Interim Report. Journal of Interdisciplinary Studies in History and Archaeology 2(1): 63-92.

BRONK RAMSEY, C. 1995. Radiocarbon calibration and analysis of stratigraphy: The OxCal program. Radiocarbon 37(2): 425-30.

--2000. Comment on 'The Use of Bayesian Statistics for (14) C dates of chronologically ordered samples: a critical analysis'. Radiocarbon 42 (2): 199-202.

--2001. Development of the radiocarbon calibration program OxCal. Radiocarbon 43 (2A): 355-63.

--2003. OxCal version 3.9. Computer Program. BRUBAKER, R. 2001. Aspects of mortuary variability in the South Indian Iron Age. Bulletin of the Deccan College Post-Graduate and Research Institute 60-61: 253-302.

BRUMM, A., N. BOIVIN & R. FULLAGER. 2006. Signs of life: engraved stone artefacts from Neolithic south India. Cambridge Archaeological Journal 16(2): 165-90.

BUCK, C.E., J.B. KENWORTHY, C.D. LITTON & A.F.M. SMITH. 1991. Combining archaeological and radiocarbon information: a Bayesian approach to calibration. Antiquity 65:808-21.

BUCK, C.E., C.D. LITTON & A.F.M. SMITH. 1992. Calibration of Radiocarbon Results Pertaining to Related Archaeological Events. Journal of Archaeological Science 19 (5): 497-512.

CASSON, L. 1989. The Periplus Maris Erythraei. Text with Introduction, Translation, and Commentary. Princeton: Princeton University Press.

DEVERAJ, D.V., J.G. SHAFFER, C.S. PATIL & BALASUBRAMANYA. 1995. The Watgal Excavations: an Interim Report. Man and Environment 20: 57-74.

DUFRESNE, A.S., J.G. SHAFFER, M.L. SHIVASHANKAR & BALASUBRAMANYA. 1998. A Preliminary Analysis of Microblades, Blade Cores and Lunates from Watgal: A Southern Neolithic Site. Man and Environment XXIII(2): 17-44.

FULLER, D.Q 2003a. An agricultural perspective on Dravidian historical linguistics: archaeological crop packages, livestock and Dravidian crop vocabulary, in E Bellwood & C. Renfrew (ed.) Examining the farming/Language dispersal hypothesis (McDonald Institute Monographs): 191-214. Cambridge: McDonald Institute for Archaeological Research.

--2003b. Indus and Non-Indus agricultural traditions: local developments and crop adoptions on the Indian peninsula, in S.A. Weber & W.R. Belcher (ed.) Indus Ethnobiology. New Perspectives from the Field: 343-96. Lanham (MD): Lexington.

--2003c. African crops in prehistoric South Asia: a critical review, in K. Neumann, A. Butler & S. Kahlheber (ed.) Food, Fuel and Fields (Progress in African Archaeobotany, Africa Praehistorica 15): 239-71. Koln: Heinrich-Barth Institut.

--2005. Ceramics, seeds and culinary change in prehistoric India. Antiquity 79:761-77.

FULLER, D.Q & M. MADELLA. 2001. Issues in Harappan Archaeobotany: Retrospect and Prospect, in S. Settar & R. Korisettar (ed.) Indian Archaeology in Retrospect (Volume II Protohistory): 317-90. New Delhi: Manohar.

FULLER, D.Q & R. KORISETTAR. 2004. The Vegetational Context of Early Agriculture in South India. Man and Environment 29 (1): 7-27.

FULLER, D.Q, R. KORISETTAR & P.C. VENKATASUBBALAH. 2001. Southern Neolithic Cultivation Systems: A Reconstruction based on Archaeobotanical Evidence. South Asian Studies 17: 171-87.

FULLER, D.Q, R. KORISETTAR, P.C. VENKATASUBBAIAH & M.K. JONES. 2004. Early plant domestications in southern India: some preliminary archaeobotanical results. Vegetation History and Archaeobotany 13: 115-29.

JOHANSEN, P.G. 2004. Landscape, monumental architecture, and ritual: a reconsideration of the South Indian ashmounds. Journal of Anthropological Archaeology 23: 309-30.

KORISETTAR, R. 2004. Origins of plant agriculture in South India, in H.P. Ray & C.M. Sinopoli (ed.) Archaeology as History in Early South Asia: 162-84. New Delhi: Aryan Books International.

KORISETTAR, R., P.C. VENKATASUBBAIAH & D.Q. FULLER. 2001a. Brahmagiri and Beyond: the Archaeology of the Southern Neolithic, in S. Settar & R. Korisettar (ed.) Indian Archaeology in Retrospect (Volume I, Prehistory): 151-238. New Delhi: Manohar.

KORISETTAR, R., P.P. JOGLEKAR, D.Q. FULLER & P.C. VENKATASUBBAIAH. 2001b. Archaeological re-investigation and archaeozoology of seven Southern Neolithic sites in Karnataka and Andhra Pradesh. Man and Environment 26: 47-66.

LESHNIK, L.S. 1974. South Indian 'Megalithic' Burials: the Pandukal Complex. Wiesbaden: Franz Steiner.

MCINTOSH, J.R. 1985. Dating the South Indian Megaliths, in J. Schotsmans & M. Taddei (ed.) South Asian Archaeology 1983: 467-93. Naples: Istituto Universitario Orientale.

MOHANTY, R.K. & V. SELVAKUMAR. 2001. The Archaeology of the Megaliths in India: 1947-1997, in S. Settar & R. Korisettar (ed.) Indian Archaeology in Retrospect (Volume I, Prehistory): 313-51. New Delhi: Manohar.

MOORTI, U.S. 1994. Megalithic Culture of South India. Socio-economic Perspectives. Varanasi: Ganga Kaveri Publishing House.

NAGARAJA RAO, M.S. 1971. Protohistoric Cultures of the Tungabhadra Valley. Dharwad: Nagaraja Rao (republished in 1984, New Delhi: Swati Publications).

PADDAYYA, K. 1973. Investigations into the Neolithic Culture of the Shorapur Doab, South India. Leiden: Brill.

--1991-92. The Ashmounds of South India: Fresh Evidence and Possible Implications. Bulletin of the Deccan College Post-Graduate and Research Institute 51-52: 573-626.

--1993a. Ashmound Investigations at Budihal, Gulbarga District, Karnataka. Man and Environment 18: 57-87.

--1993b. Further Field Investigations at Budihal. Bulletin of the Deccan College Post-Graduate and Research Institute 53: 277-322.

--1998. Evidence of Neolithic Cattle-Penning at Budihal, Gulbarga District, Karnataka. South Asian Studies 14: 141-53.

--2000-01. The problem of ashmounds of Southern Deccan in light of Budihal excavations, Karnataka. Bulletin of the Deccan College Post-Graduate and Research Institute 60-61 : 189-225.

--2002. The problem of ashmounds of Southern Deccan in light of recent research, in K. Paddayya (ed.) Recent Studies in Indian Archaeology: 81-111. New Delhi: Munshiram Manoharlal.

PADDAYYA, K., P.K. THOMAS & P.P. JOGLEKAR. 1995. A Neolithic Butchering Floor from Budihal, Gulbarga District, Karnataka. Man and Environment 20 (2): 23-31.

POSSEHL, G. & P. RISSMAN. 1992. The chronology of prehistoric India from earliest times to the Iron Age, in R.W. Ehrich (ed.) Chronologies in Old Worm archaeology (Vol. 1): 465-90; (Vol. 2): 447-74. Chicago: University of Chicago Press.

SINGHVI, A.K., D.P. AGRAWAL & K.S.V. NAMBI. 1991. Thermoluminesence Dating: an Update on Application in Indian Archaeology, in S.R. Rao (ed.) Recent Advances in Marine Archaeology: 173-80. Dona Paula, Goa: National Institute of Oceanography.

SOUTHWORTH, F.A. 2005. Linguistic Archaeology of South Asia. London: Routledge.

SUBBARAO, B. 1948. Stone Age Cultures of Bellary. Poona: Deccan College.

SUNDARA, A. 1975. The Early Chamber Tombs of South India: a Study of Iron Age Megalithic Monuments of North Karnataka. New Delhi: University Publishers.

VENKATASUBBAIAH, P.C., S.J. PAWANKAR & P.P. JOGLEKAR. 1992. Neolithic Faunal Remains from the Central Pennar Basin, Cuddapah District, Andhra Pradesh. Man and Environment 17 (1): 55-59.

ZEIDLER, J.A., C.E. BUCK, C.D. LITTON. 1998. Integration of Archaeological Phase Information and Radiocarbon Results from the Jama River Valley, Ecuador: A Bayesian Approach. Latin American Antiquity 9(2): 160-79.

ZOHARY, D. & M. HOPF. 2000. Domestication of Plants in the Old World (third edition). Oxford: Oxford University Press.

Dorian Q Fuller (1), Nicole Boivin (2) & Ravi Korisettar (3)

(1) Institute of Archaeology, University College London, 31-34 Gordon Square, London WC1H OPY, UK (Email: [email protected])

(2) Leverhulme Centre for Human Evolutionary Studies, University of Cambridge, UK

(3) Department of Ancient History & Archaeology, Karnatak University, India

Table 1. A revised chronological framework for the Southern
Neolithic, with major trends in archaeological evidence
indicated. Roman numerals indicate the Periods of Allchin &
Allchin (1968; 1982), while letters indicate new subdivisions.

 Site types,
 settlement
 pattern, Geographic Economic
Period examples distribution evidence
3000 BC Earliest Shorapur No clear
Neolithic Neolithic and Raichur. evidence of
I.A occupations, animal herding
 with ceramics. or plant
 No ashmounds. economy.
 e.g. Watgal,
 Kodekal,
 Utnur.
2500 First Shorapur, Bone
Neolithic ashmounds, Raichur, evidence
I.B e.g. Utnur, Bellary(?), for cattle,
 Budihal, Chitradurga, sheep, goats.
 Palavoy, Anatapur. No
 Brahmagiri A(?), archaeobotanical
 Kudatini(?). data, but
 Early hilltop inferred
 ashmounds in beginnings of
 Bellary cultivation
 District, system likely to
 e.g. Kurugodu, be established.
 Choudammagudda(?).
2200 Fewer ashmounds Beginnings Animal herding.
Neolithic (?). Village of Neolithic Probable
II.A sites beyond cultivation
 on hilltops. ashmound based on
 e.g. Budihal zone: native crops.
 Layer 3 village, southern
 Banahalli, T. Karnataka,
 Narsipur. northeast
 Tamil Nadu.

2000 Hilltop Beginnings Abundant
Neolithic ashmounds of villages archaeobotanical
II.B that become on Upper evidence for
 villages Tungabhadra cultivation:
 founded, e.g. River. native crops,
 Sannarachamma, plus wheat and
 Hiregudda. barley; abundant
 Hallur founded. bone evidence.
 Payaimpalli.

1800 Village Neolithic in Reports of
Neolithic continuity. Kunderu Basin chicken bone
III Sannarachamma and Cuddapah from several
 and Hiregudda District. sites. First
 villages. Greatest evidence for
 Possible number and crops of African
 subdivision density of origin c. 1500
 indicated by Neolithic BC. Possible
 Tekkalakota sites beginnings of
 Phases 1/11. (equivalent arboriculture,
 to Malwa/ fibre crops and
 early Jorwe textile
 of northern production.
 Peninsula). Copper and gold
 objects.
1400 Village Megaliths Wheelmade
Megalithic continuity, in eastern ceramics.
Transition some hilltops Karnataka. Specialised
(Pre-Iron abandoned. By end stone axe
Megalithic) Last ashmound of period, workshops. A
800 formations megaliths few possible
 cease (e.g. in wider iron implements
 Velpumudugu). region of from this period
 Megalithic Tamil Nadu, (?). Possible
 pottery and eastern finds of horse.
 burials begin. Maharashtra.

800 Classic All hilltop Megalithic Clear
Megalithic villages burials attestation of
(Iron Age) abandoned. widespread, iron working.
300 including Clear
 inland attestation
 southern of horses.
 Tamil Nadu. Earliest
 finds of
 cultivated
 rice in
 South India
 (Veerapuram).
300 Late Settlement Megalithic Rice
Megalithic/ mounds on burials agriculture
Early plains. continue, more widely
Historic and cease adopted.
AD 100 during this
 period (?).
 First
 agricultural
 village sites
 in inland
 southern
 Tamil Nadu.

Table 2. Site life-histories of ashmound sites, if
available, including dating evidence discussed in
this paper.

 Pre-ashmound Ashmound
Site evidence duration
Utnur Pre-ashmound 2500-2300
 occupation; BC (c. 200
 2800-2500 BC. years).
Kodekal Pre-ashmound Unknown.
 occupation;
 3000-2500 BC.
Palavoy None Ashmound(s),
 reported. including
 multiple
 episodes:
 2500
 BC-1700 BC
 (c. 700
 years).
Kudatini None Long period
 reported. (Phases I
 to III of
 Allchin; c.
 500-700
 years)
Budihal Pre- Ashmound 1:
 ashmound 2300-2200 BC
 occupation (c. 100
 from years).
 2400 BC.
Terdal None 2200-2000 BC
 reported. (c. 200
 years).
Sannarachamma Pre-ceramic 1900-1700 BC
 occupation; (c. 200
 minimal years).
 ceramic/
 Neolithic.
Hiregudda None. 1900(?)-1700
 BC (c. 200
 years).

Velpumadugu Unknown. 1350-1250 BC
 (c. 100
 years).

 Post-
 ashmound Megalithic
Site use use
Utnur Abandoned. None
 reported.
Kodekal Abandoned. None
 reported.
Palavoy None None
 reported. reported.
Kudatini Abandoned? Re-used in
 Megalithic
 for
 burials.
Budihal Village: None
 2300-1700 reported.
 BC.
Terdal None Re-used in
 reported. Megalithic
 for
 burial(s).
Sannarachamma Village: Village
 1700-1200/ abandoned
 1000(?) BC. by classic
 Megalithic.
Hiregudda Village: Reused as
 1700-1500 stone axe
 BC. workshop
 c.
 1400-1250 BC.
Velpumadugu Village: Village
 1250-1000 abandoned
 (?) BC. by classic
 Megalithic.