‘Afterword’:
Although the developments in Indian archaeology since this book was written in 1997–8 do not necessitate any basic or even appreciable change in its structure, argument, and contents, their implications have to be understood and put in their proper contexts.
INTRODUCTORY
The development of archaeological research in the subcontinent has gained some academic attention, leading to a better idea of the details.1 One also notes a concern with the socio-political dimensions of archaeology in modern India.2 The geographical approach to the subcontinental archaeology has not yet attracted the attention it deserves, and one hopes that with the increasing attention to the grassroots details of routes,3 etc., there will be a better reshaping of our relevant archaeo-geographical concepts. As far as the basic classificatory frameworks of the Indian people and languages are concerned, there has been no major development except the increasing frequency of DNA studies of various population groups. They throw light on the molecular make-up and possible links of various populations. For instance, in a study published in 2000, Kivisild et al., argued that India could be ‘a part of the common gene pool ancestral to the diversity of human maternal lineages in Europe’.4 In 2006 Sahoo et al.,5 observed that the Y-chromosomal data consistently suggested a ‘largely South Asian origin for Indian caste communities’ and therefore argued ‘against any major influx, from regions north and west of India, of people associated either with the development of agriculture or the spread of the Indo-Aryan language family’.
The literature of DNA studies on the Indian tribal and caste populations is fast increasing, although not yet consolidating into a range of clear historical conclusions. No prehistoric DNA has yet been isolated in India. So far the only past DNA isolated in India has been on the ad 9th century human bodies buried in the Rupkund glacier of the Garhwal Himalayas.6 In this context it is important to remember that the biological ingredients and ancestries of human population groups do not have specific archaeological or historical correlates.
THE HATHNORA FOSSIL FINDS
‘Hominin’ has recently replaced the earlier term ‘hominids’ in the case of the early human fossils. After its initial discovery in 1982, the hominin-yielding locality of Hathnora yielded from the same deposit a partial left clavicle and a left ninth rib. Both these pieces and the earlier finds have been supposed to represent a single individual, possibly a 25–35 year old female who was 135 cm (4 feet 4 inches) tall and had a modern pigmy body-frame characterised by a very robust body build.7 A further analysis ascribed8 the Hathnora skull cap to the lower part of the Middle Pleistocene beginning around 500,000 years Before Present (BP). If accepted, this would put the Hathnora find in the Home erectus category and not in the category of Archaic Homo sapiens, as thought earlier.
However, in the case of such old specimens, scientific unanimity is difficult to achieve. Some recent researches have led to specific insights into the depositional context of the Hatnora specimens.9 The geological profile of the area is currently the succession of the following ‘Formations’ (beginning at the bottom): Pilikarar, Dhansi, Surajkund, Baneta, Hirdepur, Bauras, and Ramnagar. The Dhansi Formation is placed between late Pliocene and early Pleistocene epochs. The reported vegetational sequence of the area10 begins with the Surajkund Formation. It shows sparse grassland with vegetation indicative of arid climatic conditions. The subsequent Baneta Formation, roughly 24000–20000 BP, denotes a landscape with open vegetation and the presence of waterbodies, which may indicate a cool dry climatic regime. At 8740 plus/minus 450 Yrs BP there was dry deciduous forest in the area, which suggests climatic amelioration. The basic conclusion is that in the early middle Pleistocene, the climate was warm with intermittent arid—humid phases. The terminal part of the Pleistocene was drier. Three mammalian teeth stratigraphically associated with the Hathnora fossil finds were dated by the Electron Spin Resonance (ESR) method and showed the minimum possible age for this level at 49 plus/minus 1 thousand years ago (Ka). However, these teeth could have been reworked and deposited into the Hathnora bed at any time after 160 Ka or earlier.11 Thus, many of the present conclusions about the Hathnora hominin Find are tentative. In view of the Pliocene to early Pleistocene dating of the Dhansi Formation, a remarkably interesting find is that of a few in situ Palaeolithic artefacts (generally in fresh condition) and a highly weathered fossil herbivore tooth from a thin rubble horizon at the bottom of this Formation.12
PALAEOLITHIC AND MESOLITHIC
M. Salim, a Pakistani archaeologist, has found stone tools in the Gurha Sahan area of Rawalpindi in the Upper Siwalik Conglomerate and the underlying Pinjore Formation.13 The latter has been put around 2 million years ago (mya) by the study of its magnetic polarity stratigraphy,14 and thus there need not be any doubt about the 2 mya occurrence of the first stone tools in the Siwalik zone. The claim of 1 mya dating of the Isampur Palaeolithic quarrying site in the Hunsgi-Baichbal valley of Karnataka is disputed,15 though this quarry site where limestone blocks were quarried and carried some distance by the makers of Palaeolithic tools remains a unique site in Indian archaeology.16
It is unique in the sense that for the first time a close study could be made here between the raw materials, which in this case, are slabs or chunks of limestone, and the manufacturing process of the tools. A degree of forethought and decision-making has been clear in this process.17
The basic distribution pattern of the Palaeolithic industry remains more or less unchanged. The southernmost occurrence has gone as far south as Tanjore, and in the north-east a fossil wood industry, perhaps similar to the fossil wood industry found already in Tripura and Bangladesh, has been located in Mizoram.18 This fossil wood industry is upper Palaeolithic and has close analogues with the upper Palaeolithic Anyathian industry near Mandalay in Myanmar.19 Lower Palaeolithic tools have also been reported from the Palghat district of Kerala, although one does not yet know if they form a distinct stratigraphic level or are stray finds of artefacts. The information on the environmental conditions of the Indian lower Palaeolithic still remains inadequate, and despite a focussed multi-disciplinary research programme at the site, the Attirampakkam Palaeolithic industry of the Cortalayar valley near Chennai has not told us anything historically significant except that the site was originally located near the banks of the Cortalayar and that the tools were used and discarded there to be subsequently covered by a flood of the river.20 There has also been no breakthrough in the chronology of the site. The ESR dates of its three fossil teeth are not earlier than 50/51 Ka, which possibly means that these teeth, obviously of a much later period, subsequently got mixed up with the bed containing the Palaeoliths.
It is increasingly clear that there are many uncertainties about what can be inferred from the deposits containing early Palaeoliths even through scientific research. Geological and geomorphological informations which fill up sizeable portions of modern Palaeolithic research publications in India seldom give us any dependable historical conclusion.
At 74000 years BP, the Toba volcanic eruption of Sumatra was strong enough to lay thick ash deposits in Peninsular India. Such deposits have been traced in the eastern part of the Vindhyas, the geologically old western part of West Bengal, Karnataka, and Maharashtra. The geochemical composition of these deposits tallies strongly with that of the Toba ash. A ‘catastrophe theory’ got built up in the scientific literature around this event, on the basis of the proposition that much of the earlier life was seriously disrupted by the calamity of the Toba eruption. A recent study based on the middle Palaeolithic profiles shows steady artefactual continuity between the pre-ash and post-ash deposits in India and thus undermines the ‘catastrophe theory’.21 Much of the present conclusions about the arrival of Anatomically Modern Humans in India from Africa is tentative. The event could have taken place sometime in 70000–50000 BC or even later, but the continuity of the Indian lithic sequence still makes it impossible to identify specific fresh arrivals from Africa. The decorations on the ostrich eggshell beads at Patne, which have been invoked by some as an African correlate, do not really imply anything like that because the Patne sequence from the middle Palaeolithic to the upper Palaeolithic and Mesolithic is uninterrupted and continuous. In the name of Palaeolithic research much scientific posturing has begun to take place in the modern Indian context,22 making confusion worse confounded.
At Kalpi on the right bank of the Yamuna on the way to Kanpur from Jhansi, three ‘events’ have been isolated in a dated prehistoric profile, with the base of ‘event I’ being a sand deposit of 76 plus/minus 11 Ka. Thiswasa phase of moderate to dry climate. The interface of the ‘event I’ and ‘event II’ (69 plus/minus 13 Ka) contains quartzite flaked pebbles and worked bone artefacts along with a large elephant tusk and other mammalian remains. The climate of ‘event II’ is said to be humid and its mid-level is dated 45 plus/minus 9 Ka. The basal section of ‘event III’ (humid in its lower part and dry in the upper part) is dated 43 plus/minus 7 Ka. In view of the location of the site in the alluvium at the western edge of the Ganga plain which is otherwise considered outside the belt of prehistoric finds, the Kalpi discovery23 is of great significance. The fact that bone artefacts have been found here in the middle Palaeolithic lends this site added significance.
A major upper Palaeolithic discovery has taken place at Ango Patkai in the Northwest Frontier area of Pakistan where a Sanghao-type cave site has been found to contain both stone and bone tools.24 The finds of bone tools in the Palaeolithic assemblage of the Kurnool caves of Andhra, the Kalpi profile in Uttar Pradesh (UP) and the rock-shelter site of Ango-Patkai of Bajaur clearly indicate that the occurrence of bone tools in the sub-continental Palaeolithic is much more widespread than previously thought.
Except the excavations at Jwalapuram in the Kurnool district of Andhra, which have established a sequence of Mesolithic industries between c.35000 and c.12000 calibrated years BP in association with four cranial vault fragments, an isolated tooth, pieces of red ochre used for rock painting and stone beads, there has been no major Mesolithic excavation in the past decade. Dates of the 7th to the 10th millennia BC have been obtained from the sites of Lekhahia, Mahadaha, and Damdama. These show that there need not be any doubt about the early status of the Mesolithic in the Gangetic plain and the Vindhyan fringe.25
Regarding the other Stone Age discoveries, one must mention the work on the rock-shelter site of Daraki–Chattan near Mandasore in Madhya Pradesh. Excavations have shown that the context of the many cupules found on the rock-shelter walls is a transitional phase between the lower and the middle Palaeolithic, proving that cupules have an ancestry stretching back to the Acheulian in India. Further, two engraved lines on a boulder have also been found to date from the Acheulian phase.26 The Stone Age research in the subcontinent has amply underlined the tremendous chronological span of the hunting–gathering tradition which can still be observed in various areas, and from the historical point of view this research offers insights into how the landscape was used by hunter–gatherers in the prehistoric context.
EARLY AGRICULTURE
Developments of fundamental importance have taken place in the field of early agricultural beginnings in India. Two areas have emerged into limelight: the central Ganga plain and its Vindhyan fringe, and the Aravaili system. That this problem is much more than the issue of tracing the beginning of early wheat and barley agriculture in the subcontinent in Baluchistan and then tracing its development in what is occasionally characterized as the greater Indus valley and linking it to the development of the Indus civilization, is highlighted by some recent work in the Kotumsar and Dandak caves of the Kanger valley national park of Bastar.27 Pottery, tool, or hearth could not be identified in the interior of these dark and generally inaccessible caves, but traces of fire have been found on the floors of both of them. Inside Kotumsar, about 100 m from the cave mouth, four carbon-rich samples were collected from a 32 cm thick section. Three such samples were collected from a 14 cm thick section in Dandak, about 170 m inside from the cave mouth. These carbon-rich samples have yielded dates and plant remains. It has been found that these caves were occupied or used by human groups in 4990–2080 (calibrated) BC. The plant remains indicate three types of grasses and two types of millets. One of these grass varieties is locally called Chenchrus which are still cultivated in some areas as fodder. Another variety is known as Sirwari whose leaves and tender shoots are still used as food. The identified millets— foxtail millet and another millet of the Setaria group—are still collected in the area as edible foods.
The Kotumsar and Dandak cave evidence of early plant collection is no doubt indicative of a wide tradition of such plant usage in Indian late prehistory as a background to the more unequivocal evidence of agriculture. The phytoliths of wild rice, which have been identified around the site of Lohuradeva in the Gorakhpur area at about 8250 (calibrated) BC may also indicate such a tradition of collecting wild grains as a prelude to full-fledged agriculture.
The 8th–6th millennia BC evidence of domesticated rice at Koldihawa in the Vindhyan fringe of the central Ganga plain was published in 1980. Chopani Mando of the same area yielded in its ‘advanced Mesolithic’ or ‘proto-Neolithic’ level wild rice, thus suggesting the possibility of rice cultivation in this part of the Vindhyas. Kunjhun in the Sidhi district of the eastern Vindhyas, showed the use of both domesticated and wild rice in the 4th millennium BC. The evidence of domesticated rice has been claimed for the 8th millennium BC at Damdama near Pratapgarh in the plain. The 8th to the 6th millennia BC dates have sporadically emerged for the Neolithic levels of sites like Jhusi (on the left bank of the Ganga in Allahabad) and Tokwa (the Mirzapur belt of UP). Thus, there have long been clear indications that the central Ganga plain and the adjacent upland of the eastern Vindhyas constitute an early and independent centre of the beginning of rice agriculture.
The clinching evidence has emerged from Lohuradeva (220m × 140 m × 4 m), a village on the bank of a tal or marshy depression in the Gorakhpur-Basti area of eastern UP. Studies of geomorphology, pollen records, phytoliths, diatom, and micro-charcoal were undertaken along with the excavations at the site. The presence of micro-charcoal in the 2.8 m thick succession of sediments in the marshy depression, on the bank of which the site is located, suggests slash-and-burn cultivation around 10000 BP. This is supported by the evidence of cerealia pollen and micro-charcoal in the sediments of Sanai Tal in the Lucknow area since 15000 BP. The related evidence from Dadupur (near Lucknow) and Bateswar (near Agra) attests ‘human activity’ in a larger area of the Ganga plain during the late Pleistocene and early Holocene. The domesticated rice found in the lowest archaeological deposit of Lohuradeva has been dated 8259 BP (calibrated) or 6309 BC. The phyoliths and cerealia pollens in the lake sediments indicate well-developed agriculture in the area in the 7th–6th millennia BC.
Period 1 of Lohuradeva has been divided into two phases. Period IA represents a 45 cm thick kankar-mixed deposit. The red ware and black-and-red ware, the two main types of pottery (all handmade) in this level are ill-fired, possessing coarse and porous surface and uneven core. The clay used for this pottery is mixed with rice husks and straw and not well-levigated. About half the pottery of this level is black-and-red ware, the inner core of which is about two-third black and the rest red. Both the pottery types bear slips on the inside and outside. The slip in the case of the black-and-red ware is black while the slip in the case of the red ware is red. The slip on the red ware is very fine in some cases. There are cord impressions including rope designs, on the red pottery. The pottery shapes common to both black-and-red ware and red wares are convex-sided bowls, pedestalled bowls, knobbed vessels, and miniature bowls and vases.
Burnt clay chunks and nodules with reed impressions show that the houses were made of wattle-and-daub. There are traces of circular mud-floor, post-holes, and hearths. An irregularly running channel (23 to 53 cm broad but the depth is still uncertain) has been traced through the settlement for about 12 m.
Layers of compact yellowish earth mark Period IB. The earlier pottery types continue to occur, but there are new additions including grey and black wares. The pottery of this level seems to be finer in comparison with that of the earlier level. Only a single sherd of painted pottery (a linear design in creamy white colour on a fine dark red—slipped surface) has been found. Beaker, perforated and legged vessels, and spouted vessels occur as new pottery shapes. The houses continued to be made of wattle-and-daub and there are mud floors with hearths. A terracotta sling ball, a bone arrowhead, steatite beads, animal bones showing cut marks, and plant remains constitute the other excavated material. Period IB has been dated in the late-third and early-second millennia BC (calibrated).
The climatic sequence of Lohuradeva began with a cool and dry climate in which the vegetation comprised grass interspersed by few trees and thickets. This lasted at least up to 6750 calibrated BC. There was an increase in rainfall between 6750 BC and 3750 BC (calibrated), resulting in an increased vegetation in the area. After about 3750 BC (calibrated) there was vegetation with patchy forests of Mahua and Sal trees. The use of copper and iron is attested around 2050 BC and 1250 BC respectively.
The fully domesticated rice in the lowest level of Lohuradeva occurs in association with both wild rice (Oryza rupifogon) and foxtail millet (Setaria sp.). They occur intermittently in several layers in association with other wild seeds and fruits. Barley appears at the site around 2700 BC(calibrated) and wheat a little after that. Rice was present throughout.28 After a long spell of research beginning with the excavations at Koldihawa, Chopani Mando and other sites, Lohuradeva has provided the clinching evidence of the independent and early beginning of rice cultivation in this part of the subcontinent. Its ramifications in the other areas of the Ganga plain still remain to be traced.
The claim of the Aravalli system as an early and independent centre of agricultural origin is getting increasingly strong. First, we may consider Bagor in Bhilwara. The fact that cattle, sheep, and goat of Period I at Bagor were domesticated and the fact that the earliest chronological point of this period falls in the 6th millennium BC suggest the possibility of Bagor I having an agricultural component. In the next period Bagor yielded copper implements, of which the arrowhead is identical with the arrowhead type of the subsequently excavated site of Ganeshwar (located between Kot Putli and Nimka Thana), located further up the Aravallis. In its first phase Ganeshwar was exclusively marked by microliths but possessed a number of copper tools in the next phase itself. It has not been generally realized that Bagor and Ganeshwar in two different sections of the Aravallis have the same archaeological sequence. It is likely that Ganeshwar I and Bagor I belonged to the same period, i.e., the 6th millennium BC and Ganeshwar also had an agricultural dimension like the latter site. The 6th millennium BC date for the so-called Mesolithic level in the Aravallis has also been highlighted by the mid-6th millennium BC date for a 60 cm thick Mesolithic deposit on the eastern face of Gilund-2 (Bhilwara district where Bagor is also situated). The issue needs further research, but what is intriguing is that the beginning of the chalcolithic occupation at Balathal in the same region has now been found to be about 3700 BC (calibrated). This implies a considerably earlier beginning of agriculture in the region, and as there is no reason to infer that this beginning was due to an infiltration from another area, one has to accept that this suggests an independent beginning of agriculture and metallurgy in the Aravalli zone.29
On the whole, it appears that the subcontinent had clearly developed early pockets of agriculture based on different types of cereal crops: the tradition based on the cultivation of wheat and barley around 7000 BC in Baluchistan, and the tradition based on the cultivation of rice in the central Ganga plain and the adjacent uplands at about the same period. The crop situation in the presumably early development of agriculture in the Aravallis is still unknown. The north-east, a distinct agricultural zone of the country where terrace agriculture is practised widely along with slash-and-burn agriculture, has not yet been researched at all from this point of view, but may be important in this regard in future. The situation in the southern peninsula is uncertain, there being no reason yet to push the date of the beginning of its agriculture beyond c. 3000 BC. It is also probable that throughout the subcontinent, there was a deeply embedded tradition of the collection and use of a wide variety of wild plant products including seeds of millets, as we have found in the Kotumsar and Dandak caves of Bastar.
It is historically important to appreciate the point of multiple agricultural origins in different parts of the subcontinent. For instance, the Indus civilization has so far been viewed primarily as the product of a straight arrow-line of development of wheat-barley agriculture originating earlier in Baluchistan. In view of the early rice-based agricultural centre in the central Ganga plain and the probable independent growth of agriculture and animal domestication in the Aravallis which might also include an extension in Gujarat, contributions of different strands of agriculture to the Indus civilization are more than likely. It has long been known that this civilization was based on a number of major crops—not merely wheat and barley but also rice and millets, and it is somewhat regrettable that not enough attention has yet been paid to the issue of its agricultural diversity and the impact of this diversity on its origin and development.
THE INDUS CIVILIZATION
Among the recent discoveries in the north-western part of this civilization, firm archaeological sequence has been obtained from Nal in Baluchistan (the website of Deutsches Archaologisches Institut) and the Gomal plain of the North-western Frontier region.30 Nal is a well-known site but till now there were uncertainties about its sequence and chronology. In Period I when it is found related to the Togau level of the neighbouring Kalat plateau, there was a cemetery with fractional secondary burials. The individual graves contained the remains of up to 16 individuals of all age groups, and plenty of grave goods, mostly pottery and beads of semi-precious stones and shell. Period II is the classic Nal period, although apart from the polychrome ‘Nal’ ware, there were many types of ordinary domestic pottery as well. The dead were put in single graves with a few pots. Its mud-brick houses had small rooms provided with storage bins, numerous pots, grinding stones, stone and bone implements, terracotta bull figurines, and semi-precious stone beads. Period III (6 m thick cultural deposit) was found to have close connections with central and north Baluchistan and the sites of Mundigak in south Afghanistan and Shahri–Sokhta in south-eastern Iran. There is no ‘Nal’ pottery, and in the large rooms with clay-plastered floors, gravel foundations and the use of wood in both the foundations and the roof, there were stone and bone implements, pottery, beads, terracotta figurines, and some copper seals. Copper and possibly silver were processed at the site where potters kilns have also been found. Period IV, the last phase of occupation at the site, represented the Kulli-Harappa complex. Palaeobotanical and palaeozoological studies do not denote any basic climatic change, and for the desertion of the site, some local factors like the shifting of the water-courses may be responsible. The entire 13 m thick Nal sequence has been placed between 3800 and 2200 BC (calibrated).
In the Gomal plains of the North-western Frontier, excavations at Gandi Umar Khan have led to the establishment of the following sequence in the Gomal valley: aceramic Neolithic, ceramic Neolithic (c.4500–3500 BC), Tochi-Gomal phase (c.3000–2700 BC), the transitional phase to the Kot Dijian phase, the Kot Dijian phase, and (after an occupational break) the mature Harappan phase. An interesting aspect is that both the Kot Diji phase and the mature Harappan phase have been subdivided at this site into a number of sub-phases on the basis of the occurrence/non-occurrence of particular types of pottery. The radiocarbon chronology makes the Kot Diji phase of this region late in comparison with its date elsewhere, and one now notes a reluctance among some Pakistani archaeologists to accept Kot Diji as ‘early Harappan’. However, the areas in which the Kot Diji-type of pottery have been traced are vast, and its cultural picture was not the same everywhere. The chronology also could vary significantly.
Among the sites of the Indus civilization, which have recently emerged into limelight, one may mention Kanmer, Shikarpur and Juni Karan in Gujarat, Bhirrana and Farmana in Hary ana, and Sinoli and Alamgirpur in western UP. Kanmer, near Rapar like Dholavira, is a fortified settlement with about 9.5 m high fortification wall and the evidence of rice, wheat, barley, and lentils. There are four cultural stages at the site, beginning with the early Harappan and ending with the post-Harappan. An interesting feature of the site is the find of a large number of microliths made of steatite. At the edge of a narrow creek of the Gulf of Kutch, Shikarpur is a 3.4 ha site located on the top of a sand dune. Of its 6.40 m thick occupational deposit, Phase 1 is mature Harappan, but Phase 2, although mature Harappan because of its typical pottery, Rorhi-type chert, shell bangles, stone beads, terracotta beads, bangles and triangular cakes, cart frames and wheels, and copper implements has regional pottery types. Phase 3, much eroded, occurs at the top. A fortification wall has been traced at an edge of the site. Juni Karan on the old Sind to Kutch route, located in the Khavada island of the Rann of Kutch, about 42 km to the north-west of Dholavira, is a fortified citadel complex within an outer fortification.31 The outer fortified area (220 m × 225 m) had in its northern part a citadel area (72m × 92 m). The fortifications were made of rubble set in mud mortar. The gateway in the south-west was approached by a 3.4 m wide corridor which was originally 6.42 m wide. Further east along the fortification, another gateway had a set of large chambers and led to Stadium 1 (25 m × 8.25m) which had at its eastern edge Stadium II (22.20 m × 14.50 m). Stadium I, basically a brick platform sloping away from the chambers of the gateway to the ground in front, has been supposed to be the common people's stadium, whereas Stadium II is supposed to have been meant for the elite. There is an open area of 200m × 50m between the stadia and the ‘lower town’. Juni Karan partly duplicates the planning of Dhola Vira, but the finds of two stadia and the fact that it is located in another island of the Rann of Kutch make it a very important site.
Bhirrana (150 m × 220 m) in the Fatehabad sector of Haryana, where Banawali and Kunal are located, shows four phases: Hakra ware phase (IA), early Harappan (IB), transitional to mature Harappan or early mature Harappan (IIA), and mature Harappan (IIB). The occurrence of dwelling pits which were up 3.40 m in diameter, 34 cm to 58 cm deep, and mud-plastered with occasional brick-linings on the inside, recalls similar finds at Kunal. It has been assumed by the excavator that these dwelling pits without any trace of post-holes around them could accommodate three to four persons and had in all likelihood superstructures of light material. There were also smaller pits (2 to 2.20 m diameter and 0.8 to 1.08 cm deep), some of which contained charred bovine remains with horns severed at the base, and grains. These have been inferred to be sacrificial pits. This occupational level is associated with the Kalibangan type of early Harappan pottery and a bichrome pottery (outline in black with white inside). Graffiti were observed, although rarely, on the pottery. Two copper rods, shell and terracotta bangles, and beads of semi-precious stones were among the other finds from this phase. Phase 2 was more prosperous, showing a fortified and planned settlement with rectangular mud-brick houses and streets, dish-on-stands and button-based goblets as pottery shapes and a hoard of 3461 semi-precious stone beads including 248 specimens made of lapis lazuli. The mud fortification wall of the mature Harappan phase was 2.15 m to 3.75 m wide and was accompanied by multi-roomed mud brick houses with drainage, hearths or ovens, brick-laid bathroom floors and storage pits. A copper-smelting area has been identified outside the western fortification wall. Two streets, one along the fortification wall and the other in the central part of the settlement, have been excavated. The pottery is typically mature Harappan, and among the copper tools one notes Ganeshwar-type arrowheads and two inscribed celts. Spokes, similar to the ones at Banawali and Rakhigarhi, have been found painted on terracotta wheels. Gold beads, five steatite seals, an ithyphallic terracotta, a female terracotta figurine and the engraved figure of a woman recalling the pose of the famous ‘dancing girl’ of Mohenjodaro on a potsherd are among the other important finds.
Among the pottery from Bhirrana, one notes what has been called Hakra ware on the analogy of sites in Cholistan—and the occurrence of this type of pottery which in Cholistan pre-dates the early Harappan phase at some Haryana sites including Kunal, Rakhigarhi, and Bhirrana—poses a problem regarding the current idea of development of the Indus civilization in Haryana. What is interesting is that this pottery has been found in the recent excavations at Farmana, basically an early and mature Harappan settlement site with burials near Rohtak, and further east, at Alamgirpur where the Harappan occupational level has been found interlocked with the Painted Grey Ware level, thus repeating the Bhagawanpura evidence of Haryana in the upper Doab as well. The Hakra ware level has been inferred by its excavator, the late L.S. Rao, to date from the 4th millennium BC. Whether the Hakra ware found as far east as Farmana and Alamgirpur dates from the same period remains to be seen. Sinoli near Baghpat in Haryana and not far from the left bank of the Yamuna has revealed a large burial site where the burials have been found mostly extended. The copper objects found among the grave goods are bangles, antennae-swords, and anthropomorphs. The last two types are typical Gangetic valley ‘Copper Hoard’ specimens and their occurrence in a burial of the Harappan tradition at Sinoli clearly indicates that the Copper Hoard assemblage was a part of the Harappan tradition. A copper sheath of an antennae-sword was also found. This is the first time that such a discovery has been made. The copper anthropomorphs have been found possibly in a ritual context. The finds comprise, in addition to the copper objects, semi-precious stone beads, micro-beads of steatite and a corpus of rich well-slipped plain and painted red pottery whose shapes include elongated, narrow-necked and footed vases with out-turned rims, and short and squat dish-on-stands. The date of Sinoli is uncertain but cannot be much different from the nearby excavated site of Hulas. Another major Harappan discovery in this area is that of a rich hoard of gold and silver jewellery and semi-precious stone beads at Mandi in the Muzaffarnagar district, which indicates that the upper Ganga plain or at least the northern part of the Doab had a prosperous Harappan phase which also incorporated the tradition of the Gangetic valley copper hoards.32
An isolated find from Kalibangan deserves special attention. A small terracotta Sivalinga has been located among the excavated mature Harappan antiquities from the site. The identification is indisputable and sets at rest a long-standing controversy about the presence of Siva worship in the Indus civilization.33
Some new insight has been obtained in recent years on the distribution of the Indus sites in Haryana, Indian Panjab, and western UP. First, in western UP the distribution of such sites is right up to the foothills of the Siwaliks. Painted pottery of this type has been found on the right bank of the Ganga in the shadow of the hills of Haridwar. Second, in Haryana, the broad trajectory of the Mature Harappan distribution is the alignment of Sirsa, Fatehabad, Hissar, Hansi, Bhiwani, Rohtak, and Jhajjar, with only one site—Balu—in the Jind-Kaithal stretch. This was the broad line of movement from Sirsa to the area of Delhi in the historical periods too, and here we can only observe this general similarity without trying to offer any specific explanation. At Badli, east-south-east of Jhajjar, R C Thakran34 has excavated a Mature Harappan level, and the location of this site suggests a Mature Harappan thrust towards the Yamuna. In Indian Panjab, sites of the Harappan painted pottery tradition occur in the districts of Amritsar and Gurdaspur and suggest a similar distribution in the whole of Lahore-Gujranwala-Sialkot sector.
As far as we can judge, the occurrence of sites of the Harappan painted pottery tradition in the districts of Amritsar and Gurdaspur are important in their own right, indicating their occurrence in the Lahore-Gujranwala-Sialkot sector of Pakistani Panjab. The site of Manda south of Jammu, which belongs to this tradition, also suggests the occurrence of sites of the Harappan tradition in this part of Pakistan.
The find of painted pottery sites in the Hoshiarpur area—even Nawanshahr from where painted pottery and plain red pottery of thick fabric occur at Rahon—indicates that there was a chain of sites of the Harappan tradition at the foot of the Panjab Siwaliks. Rupar and other sites of the area, including Chandigarh, may be considered part of that zone in the Siwalik foothills and the submontane areas. The painted pottery sites of the Ferozepur district fall in a border area with Pakistan, and again, one can expect such sites south of Lahore. Ferozepur itself easily leads to Pakpattan on the Pakistani side. In fact, the location of Harappa, when viewed in the light of the route which came both from Sind and the Gomal plain to join the Sirsa to Rohtak alignment, clearly shows that it was a major site on this route.
The idea that there was a Harappan migration to the east in the late phase of this civilization has been propounded quite frequently. Even without going into the archaeological nitty-gritty of this idea, one can say that the sheer fact of Harappan distribution in a massive zone going across the riverine plain to the long Siwalik belt from Jammu to Haridwar, covering Panjab, Haryana and western UP considerably nullifies this idea of the Harappan sites being comparatively late in this region. The sheer geographical expanse of their distribution makes one realize that such a close adaptation to this landscape can only imply a deep root and long developmental process for the details of this site distribution in Haryana and Indian Panjab.35
Finally, there is an increasing tendency to relate the growth and decline of the Indus civilization to the fluctuations of the monsoon.36 Whether this type of environmental determinism offers a logical way to understand the phenomenon of the Indus civilization can always be debated.
INDIA BEYOND THE INDUS CIVILIZATION
New information is available from most of the areas, but much of it is only in bits and pieces. The date-range of the Gandhara Grave Culture in Chitral has been found to be from c. 1000 BC to c. AD 1000, thus bringing this grave tradition down to the Islamic period.37 In south-east Baluchistan, there is apparently a desertion of sites after the Kulli complex, which is inexplicable. The earliest radiocarbon dates from Balathal in south-east Rajasthan suggest its beginning well inside the 5th millennium BC and strongly reinforce my idea that there was an early and independent centre of agriculture and metallurgy in the Aravallis. Ojyana is a major newly excavated site in this region. It was a fortified settlement of the Ahar culture and has yielded evidence of streets and regular planning of houses. A diversified range of food-crops included wheat, barley, rice, jowar, ragi, and Setaria millets, a variety of legumes, sesamum, and safflower. There is no radiocarbon date from the site. Gilund was excavated in 1999–2004 but did not yield anything novel except a complex with parallel walls. In fact, sites have been dug all over India up to the deep south, but only very few of them are important.38 In the deep south Adichchanallur on the bank of the Tamraparni in Tirunelveli is one such site. Claims have been made for a separate habitational area in this large urnburial site where three phases of burials have been established. The area was used for copper and irori-ore mining before it was used as a burial place. The burial urns of Phase I were buried cutting the bedrock and the bodies were placed in the crouching position along with couex beads and iron dagger, sword and knife. There were two more phases of burials after that. There are two Optically Stimulated Luminiscence (OSL) dates from the site: 1570 BC (plus/minus 570 years) and 3750 BC (plus/minus 1200 BC). The dates permit a wide range of variation, but the beginning of the site may securely be placed in the second millennium BC.39 Major discoveries regarding the beginning of iron have been made at Malhar (near Banaras) and Dadupur (near Lucknow) in UP. At Dadupur the earliest date, c.1700 BC, is estimated on the basis of three separate radiocarbon dates, whereas at Malhar the calibrated ranges of the two available dates are 1882–1639 BC and 2012–1742 BC. The Malhar finds include an iron ploughshare, thus proving that iron began to be used in agriculture, at least in the Banaras part of the Ganga plain, by the beginning of the second millennium BC.40
EARLY HISTORY
The date of c.800 BC has emerged for the earliest phase of the Northern Black Polished Ware (NBP) at several sites including Ayodhya, Jhusi, and Gotihawa. I would now put the first phase of the beginning of cities in the subcontinent between c.800 and c.500 BC. There is no reason to think that the spread of the NBP to the areas outside the Ganga plain was significantly later than the date of its beginning in the former area. For instance, the Ganga plain and the Deccan were well-linked by a complex of routes, which eventually led to the deep south. The NBP-bearing level of Wari Bateshwar in the Brahmapulra-Meghna delta has been dated in the 5th century BC, and from Korkai near Tirunelveli there is a date of c. 9th century BC for the NBP. At Alagankulam near Madurai the NBP has been found in the early context of the site, and the quality of this NBP is as good as the quality of the early NBP from the Gangetic valley. Once one accepts the argument that the NBP could easily reach the areas beyond the Ganga plain on the back of the age-old trade between the Ganga plain and these areas, we have to revise the generally late dates which have so far been inferred for the beginning of the early historic period in Andhra and the southern states of Karnataka, Tamilnadu, and Kerala. Pending further research, I would put the date of the early historic beginning in each of these areas around 500 BC.
A major early historic discovery of the recent years is the identification of Muziris, mentioned in the Roman sources, with Pattinam near Kodungallur (north of Cochin) on the coast of Kerala. Occupation began at the site in the first half of the first millennium BC and the height of the trading period was from the 1 st century BC to the AD 5th century. The site has a rich repertoire of imported West Asiatic and Roman pottery, and yielded a ‘wharf’.41
The trade network of the early historic period has been highlighted by a field-study of the routes which went from the Ganga plain to the Deccan. In the Ganga plain, the starting points were Pataliputra, Banaras, Allahabad, Kausambi, Kanpur, Mathura, and Agra. It was a complex of routes in which various areas of central India (Rewa, Malwa, the Jabalpur area, Hosangabad, Shadol-Bandhogarh, Burhanpur, Chhattisgarh, etc.) played a leading role in the sense that the routes to Maharashtra and Andhra went through them.42 One hopes that the grassroots study of such routes will throw a lot of light on the network of communications which linked the various parts of the early historic subcontinent. Many of these routes continued to be in use till the modern period.