The North-west
The main units of study are the Peshawar plains, the area to its east and north-east up to Chitral, the plains of Kohat and Bannu and the Potwar plateau. The Peshawar plains are about 6400 sq km of semicircular lowlands covered by spreads of gravelly or alluvial detritus. The Kabul river meets the Indus at the eastern edge of the plain and the Swat river meets the Kabul in the plain itself. Agriculture is largely dependent on irrigation, the annual rainfall being 160–240 cm. The main areas between Hazara–Malakand and Chitral are, from the east to the west, Hazara, Buner, Swat, Dir and Chitral. The hills cascade from the 7,580 m high Tirich Mir in Chitral down to 1500–1800 m high ranges in Malakand and other areas which skirt the plains. The ranges have a NE–SW trend and deep narrow valleys between them. The main river is the Swat (arable land on either side in narrow strips), but the Kunar comes down from Chitral and moves into Afghanistan to join the Kabul. Kohat has about 260 cm of annual rainfall, but there are springs in the limestone formations of the area. The Tochi and the Kurram are the main rivers in Bannu which has an annual rainfall of 180 cm. Wool is an important product of both Kohat and Bannu. The Potwar plateau lies between the Indus and the Salt range, with an area of 10,000–13,000 sq km and an annual rainfall of 250–400 cm. Raw materials are very sparse in the whole region: some iron and antimony in Chitral, and salt mines in the Salt range. Many trade routes pass through the region, the Khyber being the best known, but there are also other lower passes from the Kunar into Dir and Swat. The Karakoram route to central Asia through Chitral, Hunza, Gilgit and the Pamirs is well known too and shows an enormous range of rock paintings and engravings along the way, especially in the Gilgit sector. Early historical inscriptions too are many and contain in some cases such personal statements as ‘Jivachandra went this way’ or ‘there was a priest called Kirtibhushana’. North-eastern Afghanistan, the lapis lazuli country, can also be approached from Chitral. From the Potwar plateau one can travel easily to Kashmir, and from Bannu one can follow the Tochi and Kurram alignments into Afghanistan. On the Indian side the Indus system goes down to Sind, and Delhi is easily accessible along the Punjab piedmont.
The early historic sites of the north-west, which embody the long historical sequence of the area from the Achaemenids to the Sassanids and Hindu Shahis, are mostly settlement sites and Buddhist stupa and monastic sites. The latter category of sites occurs both in the plains and on the hill slopes and tops of the valleys through which the trade routes pass. Between 1915 and 1920, the Archaeological Survey of India reported 397 sites in the Peshawar area, and recently Ihsan Ali of Peshawar University found 1228 multi-layered sites in the Peshawar plains. A number of excavated stupa sites of this region are famous: Takht-i-Bahi, Shahri Bahlol and Shahji-ki-Dheri, which are all in the Peshawar plains, and Manikyala in the Potwar plateau. In Bannu the major historic site is Akra (200+ acres, possibly the capital of an Achaemenid satrapy) which yielded Indo-Greek coins. In the Peshawar plains Charsadda and Hund are the major mounds while Taxila is an important site in the Potwar plateau. The ruins of Charsadda or the ancient city of ‘Pushkalavati’ are spread over 10 sq km and two major mounds: Bala Hissar and Shaikhan Dheri (600 m north-north-west of Bala Hissar). The Bala Hissar stratigraphy begins basically with the Achaemenids, and when the Greeks under Alexander laid siege to it, the Bala Hissar citadel enclosed an area of 15 acres with a 5 m wide mud rampart and a 4.5 m wide and 3 m deep moat. The Shaikhan Dheri stratigraphy runs from the middle of the second century BC to the close of the second or the middle of the third century AD. This grid-patterned city was founded by Menander and it declined after the Kushanas. The city plan, of which three parallel streets and a side-street crossing at a right angle have been excavated, belongs to the Kushana period. The occupational area of Taxila or ancient ‘Takshasila’ is divided into three mounds—Bhir, Sirkap and Sirsukh—in three different parts of the valley and spans the sequence from the Achaemenids to the Kushanas. About 3 acres of mostly Mauryan remains have been unearthed in the middle of the Bhir mound: irregular plan, four streets, five lanes and associated house blocks. One street was 6.6 m wide and the rest 2.7–5 m. The houses had open courtyards, and in one case the ground floor had 15 to 20 rooms. The enclosing stone wall of Sirkap (to the north-east of the Bhir mound) was 5 km long, 45–6.5 m wide and 6–9 m high. It was strengthened by berms and rectangular bastions. The only excavated gateway is the northern gateway which has a hall. For a length of about 600 m the Indo-Parthian level of the city has been excavated, showing a grid plan, house-blocks interspersed by stupas and temples. One house covered 2160 sq m and the existence of a ‘palace’ has been inferred in the south-eastern extremity of the excavated remains. The third city, Sirsukh, came up towards the close of the first century AD and a masonry rampart wall with bastions enclosed a space of 1350 × 990 sq m. In the hills bordering Taxila there were many Buddhist remains, the most important of which is the Dharmarajika stupa founded by Asoka.
Kashmir, Punjab, Himachal Pradesh, Sind
Kashmir is a narrow valley hemmed in by high mountain ranges on all sides, and thus the issue of routes is important in this context. The north-eastern flank is marked by the Himalayas (the dominating peak in this sector: Nanga Parvat, 8126 m high) and the south-western by the Pir Panjal range (maximum height 500 m, length 240 km, width 105 km). Three major approaches to the valley from the plains are the Banihal pass (from Jammu), Baramula pass (from Rawalpindi) and Pir Panjal pass (from Gujrat in Pakistan). The routes from the valley to central Asia emanated from Gilgit in the west and Leh in the east. From Gilgit one goes to Yasin and from Yasin to Taghdumbash Pamir across the Darkot, Baroghil and Wakjhir passes. From this area there is a direct route to Kashgar across the Chichiklik plateau. From Yasin one can also go to Mastuj and Chitral. With Yarkand as the objective one has a choice of three passes after Hunza: Shimsal, Khunjerab and Mintaka. From Leh the target was Yarkand, 825 km by ‘summer route’ across the Karakoram pass. There is a ‘winter route’ as well through a few other passes, but the most important and easy route between Leh and Yarkand is through the Changchenmo valley. Skardu, the centre of Baltistan between Gilgit and Leh, also provided an access to central Asia through Saltoro glacier. Further, Tibet is approachable from Ladakh, mainly along the Indus. Apart from the temple ruins, the early historic archaeological evidence from Kashmir is nothing more than the stratified sequence at Semthan beginning with the NBP, the excavated apsidal temple with a rectangular antechamber at Harwan (c. AD 300), the Kushana period inscription at Khalatse bridge on the Kashmir–Ladakh route, and the ‘Dard graves’ of Ladakh (? AD 1 to 500). No site has been excavated in detail and properly published.
A recent survey in mainly the eastern part of Pakistani Punjab has identified about 130 sites dated between the first and seventh centuries AD. No details are yet published but the distribution map suggests a general concentration east of the Chenab. At Tulamba near Multan the large basic site contains an inner mound of square shape and its first two periods ranged from the sixth to fifth century BC to the third century AD. In Indian Punjab Madhu Bala gives a district-wise list of 229 sites—Amritsar: 35; Bhatinda: 32; Hoshiarpur: 12; Gurdaspur: 32; Ludhiana: 33; Patiala: 43; Ropar: 9; Sangrur: 13; and Jalandhar: 20. Detailed explorations should reveal more but as of now nothing else is known even about the published ones except their names. Rupar, located at the point where the Sutlej enters the plains, has NBP and somewhat later, a retaining wall of a large tank with an inlet to bring rain water to the tank. Sanghol in Ludhiana has a mud fortification and three concentric moats around.
Fig. 42 Early historic pottery, Rupar (‘Ancient India’, 9)
Fig. 43 Early historic pottery, Rupar (‘Ancient India’, 9)
No major early historic site has yet been excavated in modern Himachal Pradesh, although there is no doubt about the existence of such sites there. Pottery dating from c. 200 BC has been noted at Ror in Kangra. An ‘urban Kushana settlement’ has been briefly investigated at Dat Nagar in Simla and there is enough evidence in the form of early coins from Kulu and other regions.
The evidence from Sind, with the exception of Brahminabad–Mansura, is limited only to stupa sites, about seven of which are reported in the archaeological literature. The Indus delta has moved inland and has no doubt been subject to other drastic changes. In the adjacent tract of Baluchistan a few miscellaneous sites exist, among them Mastung in Kalat which yielded a double-handled silver chalice of a type reported from the Scytho–Parthian levels of Sirkap, Taxila.
The corridor from Sind to Rajasthan and Haryana is formed by the dried-up course of the Ghaggar–Hakra, and the segment with early historic sites lies on the Indian side of the border and shows more than 20 sites in a comparatively small area between Anupgarh and the Haryana border. Here the excavated site is Rangmahal whose naturalistically painted black-on-red ‘Rangmahal pottery’ has been found widely distributed in the context of the early centuries AD in the adjoining areas of Rajasthan and Haryana.
Haryana and the Upper Gangetic Valley
If one travels to Delhi from the Rajasthan–Haryana border near Sirsa one passes through a wide tract of alluvium between which and the upper Gangetic plain there is no topographical marker. There are hundreds of early historic sites in this region, but unfortunately the various district survey reports which contain the names of these sites are still lying unpublished in the Kurukshetra University. The names of three very large sites are easily recalled: Agroha, Thaneswar and Sugh. At each of them the mounds continue for several kilometres; at Thaneswar the mound supposedly measures 75 ha but may not stand for its complete extent. Except for some minor structural details and their basic stratigraphic sequence, nothing is known. The situation is not particularly better in the upper Ganga plain. There cannot be any lack of early historical sites; in Kanpur district alone their total number (c. 600 BC–c. AD 200/300) is 240. In the NBP phase there were 99 settlements, of which 65 were located on the river banks. There are 81 settlements of less than 2 ha in size. The size of only 14 settlements of this period varied between 2 and 5 ha, and only 4 NBP phase settlements were more than 5 ha in size. Out of 141 settlements of the post-NBP period, only one settlement was more than 10 ha in size. Most of them measured below 2 ha. The average spacing of settlements in the NBP phase was 9 km, and for the post-NBP phase it was 7.5 km. These data apply only to Kanpur district and may not be wholly accurate, but still offer a general insight regarding the Doab as a whole. However, the dominant sites of the region, which may be considered urban centres, are largely unknown beyond a few known sites: Hastinapur (c. 31 ha), Purana Qila in Delhi, Atranjikhera (c. 46 ha), Mathura (360 ha or more), Kampil (Farrukhabad district, UP), Sankisa (Farrukhabad district, rampart more than 5.5 km in circuit), Kanauj, Chakranagar (Etahwa district, UP, rampart circuit more than 5 km), Ahichchhatra (Bareilly district, UP, fortification 5.6 km in circuit), Saketa–Ayodhya (4 to 5 km in circuit), Hulaskhera (Lucknow district), Kausambi (the present ruins over 20.5 sq km), Bhita (Allahabad district, UP, fortified) and Sringaverapura (1 km along the Ganga, Allahabad district). The stratigraphic sequence of these and other sites of the region is known; in the case of the major sites like Mathura and Kausambi the details of their fortification system are also known. In the late centuries BC and early centuries AD, i.e. in the second phase of its existence as an early historic city, Mathura had a quadrangular inner fortification in the centre of a longish crescentic outer fortification system. High mud rampart walls (topped by brick fortifications in later phases in some cases, cf. Ahichchhatra, Kausambi), bastions, watch-towers, moats and gateways figure time and again in the archaeological record but the lack of horizontal excavations in the city area keeps the present knowledge on the most basic level. It is only in John Marshall’s excavations at Bhita that we come across streets (one of them, High Street, 9 m wide, and the other, Bastion Street, somewhat narrower) and houses along them. A Mauryan-period house on the High Street (12 rooms arranged around a courtyard and possibly with an upper storey) has been called ‘house of the guild’ on the basis of the find of a seal mentioning guild. At Sringaverapura a tank complex (dated in the late centuries BC) with a storage capacity of 2 million litres of water, which was built to channelize the water of the Ganga and to store and desilt it, is a major discovery.
However, many major sites in Haryana and the Doab possibly lie unknown or ignored. In the Bulandshahr area, Ahar (ancient ‘Tattvanandapura’?) measured 4 sq km. According to the 12th volume of the Cunningham reports, the irregular oblong mound of Indor Khera on the Kali Nadi in the Aligarh area measured about 10 ha but stood 21 m high. ‘The ancient city would seem to have been situated on an interiorly high raised earthwork platform.’ Another place in the same general area but on an old course of the Ganga, Sankara, measured about 55 ha and is simply a high steep-sided, flat-topped earthwork, or a sort of lofty table-topped plateau of earthwork which never had any kind of raised rampart whatever along its edges, but seems to have depended principally on height and steepness for means of defence.
Fig. 44 Sringaverapura tank (Lal, 1993)
The Middle Gangetic Valley and the Trans-Sarayu Plain
This region includes eastern UP and Bihar. There is a plethora of detailed exploration reports on eastern UP, either district-based or based on a small river valley. In old Basti district, which included the modern Domariyaganj and Siddharthanagar areas, there are 165 sites, many of them multi-layered and early historic. In the trans-Sarayu region as a whole, Sravasti (rampart circuit c. 8 km) is the most important site, of which very little has been excavated except monastic ruins. The Piprahwa-Ganwaria complex, proved to be ancient Kapilavastu, is a much smaller site and here too only the religious remains have been excavated. However, this region ‘could boast of a large number of smaller urban sites with an average size of 20 to 25 acres or more’. The inscriptionally known Maurya administrative centre of Sohgaura and the glass-manufacturing fortified site of Kopia, both on the Ami river, belong to this category. In modern Ballia district 12 sites showed NBP whereas 20 sites belonged to the late centuries BC–early centuries AD bracket. In the middle Gomati valley, which covers much of Sultanpur district, there is a steady progression in the number of sites from c. 600 BC to c. AD 300; in the early centuries AD the number is 84, whereas between c. 600 and c. 300 BC it is a little more than 10. Apart from Rajghat or modern Banaras, there are at least five fortified city sites in the Ballia–Banaras sector: Pakka Kot, Khairadih, Nahush-ka-Tila, Masaon Dih and Bairant. Some unfortified sites too can claim urban status because of their large size (cf. Nal-ka-Tila, Robertsganj). The beginning of fortification at Rajghat is c. 600 BC.
In the broad sweep of north Bihar, detailed survey reports are lacking, but the fortified sites are: Manjhi (a Bhita-type site, badly excavated), Katragarh (Darbhanga sector), Vaisali, Kopagarh, Balirajgarh (the last two in Madhubani), Mangalagarh (Samastipur), Naulagarh and Jai Mangala Garh (both in Begusarai). Fortification walls have been excavated at Katragarh and Balirajgarh but there has been little progress beyond them. In south Bihar, the major sites are Chausa, Buxar, Deo Umga (?), Bodh Gaya and Devangarh, in addition to the already well-known ones of Rajagriha, Pataliputra, Sonpur and Champa. Devangarh is a fortified urban site to the south of Rajgir whose outer cyclopean masonry wall on the hill tops is famous, but one cannot be sure of the relationship between ‘old’ and ‘new’ (fortified) Rajgirs. The timber pallisade and the stone-pillared hall of Pataliputra echo Megasthenes’ description of the place, but these lie in the periphery of the settlement and not in the core, which lies untouched. Champa (the Nathnagar locality of modern Bhagalpur), the capital of Anga, was a most impressive site—a high, artificially formed table-top plateau with enclosing fortification, which overlooks the long and narrow stretch of ancient occupation in the narrow strip of ground on the bank of the Ganga.
Assam and the North-east, the Padma-Bhagirathi Delta (Bangladesh–West Bengal) and Orissa
There is no positive evidence of early historic occupation as yet in the Brahmaputra valley and the north-eastern region as a whole. But this has more to do with the lack of excavations in the region than with real absence. Tezpur seems to have yielded some early pottery on the surface but there has been no systematic excavation at this site. Claims have been made for the late centuries BC level in Meghalaya and Manipur, but more work is necessary. There are two major early historic sites in Bangladesh: Mahasthangarh or ancient ‘Pundranagara’ (c. 185 ha, fortified) on the bank of the Karatoya river in northern Bangladesh and Wari-Bateshwar at the mouth of the Meghna delta. Both these sites had links with the Brahmaputra valley. In northern West Bengal, there is only one fortified early historic site, Bangarh. In the Bhagirathi delta, along an old course of the Ganga below Calcutta and along the Midnapur coast there are quite a few urban sites, the most important of which are Chandraketugarh near Barasat and Tamralipta (modem Tamluk). Most major sites have, however, remained unexcavated. This belt must have catered to both Roman trade and south-east Asian trade in the Bay of Bengal. In addition to these deltaic sites, there are a few more early historic urban sites in West Bengal: Kotasur (fortified) in Birbhum, Mangalkot on the bank of the Ajay in Burdwan, Pokharna on the southern bank of the Damodar and Dihar on the bank of the Dwarakeshwar. All these sites have been marginally excavated. In Orissa, the fortified early historic sites are Sisupalgarh (near Bhuvaneshwar) and Jaugada, further down the coast. Early historic evidence occurs elsewhere too, and a significant discovery of recent years is that of Radhanagar (brick fortification enclosing 6 sq km) on the bank of the Keluo, a tributary of the Brahmani near Jajpur. This has been identified with ‘Dantapura’, a capital city of ancient Kalinga.
Andhra, Karnataka, Kerala and Tamil Nadu
The main focus is on the Krishna–Godavari delta and the adjacent uplands which played a role in both maritime and internal trade. There are a large number of early historic Buddhist monastic sites in this region, and among the major excavated settlement sites are Nagarjunakonda among the hills on the Krishna (a fortified citadel, a residential area with a stadium, and many temples and stupas on the river bank), Satanikota on the Tungabhadra in Kurnool (rampart, moat, gate), Dhenukataka near Amaravati, and Kondapur. There are also other marginally explored and excavated sites. In Karnataka, the largest site seems to be Sannati (c. 200 ha, brick fortification, rich in Buddhist relics) but sites like Udayavara (near Udupi), Banavasi (north Kanara), Vadgaon–Madhavpur (Belgaum district) and Isila (Brahmagiri in Chitradurg) deserve attention. Isila was a Mauryan provincial centre. Roman references to coastal sites in Kerala notwithstanding, nothing early historic has yet been identified and documented in Kerala.
K. Rajan’s work in Tamil Nadu has thrown up some extremely important early historical evidence in that region in recent years. The focus is mainly on the area between the Palar and Vaigai–Tamraparni rivers. The archaeology of this region has hitherto been dominated by Roman trade, as evidenced at Arikamedu near Pondicherry, and megaliths. The antiquity of the Roman trade with this region has been pushed back to the last two centuries BC, with its peak between 50 BC and AD 50 on the basis of recent excavations at Arikamedu. Another port site of the region, Kaveripattinam, has its first period dated between the third century BC and first century AD, and during this period it had successively a wooden boat jetty and a brick-built wharf. Rajan has demonstrated that the myriads of megalithic sites in this area are associated with settlements, some of which are fairly large, up to 50 acres. He points out that the Pandya and Chera kings were issuing inscribed coins as early as the third century BC and also draws attention to the large-scale occurrence of post-firing graffiti on pottery, which, according to him, suggests a clear connection between these graffiti and the Tamil–Brahmi script which appears later. In fact, he claims that ‘in the present level of knowledge one should consider these marks as an undeciphered script.’ Rajan argues that at the excavated site of Kodumanal the Tamil–Brahmi script goes back to the third century BC and that its standardized form and grammar during this period suggests its still prior development. In this sense the mid-first-millennium BC date of the Brahmi script at Anuradhapura makes sense, and the same date may cogently be applied to the beginning of the Tamil–Brahmi script in the Kaveri delta and elsewhere in Tamil Nadu. Finally, Rajan’s excavations at Kodumanal have demonstrated the extent and richness of raw material extraction and processing (cf. Beryl, iron and steel) and thus the presence of trade, possibly well before the third century BC. Moreover, he argues for a close interaction with north India at this stage, mostly in the very presence of the Brahmi script and such items as carnelian beads which were not available locally. According to Rajan, ‘it all suggests that the foundation for the emergence of the state would have been laid down at least in the fifth century BC’ Thus, we are confronted with the distinct emergence of early history in the Kaveri delta as early as the fifth century BC.
Maharashtra, Madhya Pradesh, Gujarat and Rajasthan
It is around this period and later that Maharashtra passes into early history at places like Nasik, Brahmapuri, Paithan, Ter, Kaundanpur, Bhokardan, Adam and others. Rampart and moat exist at most of these sites. Another major site is Sopara on the Konkan coast. Many Maharashtra sites, especially those in the interior, are linked to the internal trade routes directed towards the coast for maritime trade. Apart from Malhar in Bilaspur district in the Chhattisgarh sector, the major early historic sites of Madhya Pradesh are scattered in its Malwa region: Ujjayini, Vidisa and Pawaya. Rampart at Ujjain goes back to c. 600 BC, and was 73.5 m broad at the base and 132 ft wide at the top, protected by well-cut and well-laid wooden logs (teak and Safed khair) for about 105 m along the river to prevent the bank from getting scoured by the river-action. Vidisa was fortified too, and if Ujjayini was the chief city of Avanti or western Malwa, Vidisa was the chief city of Akara or eastern Malwa. Not much is yet known about early historic Gujarat. Excavations at Broach, a port town at the mouth of the Narmada, Shamlaji (brick fortification), etc. have given an idea of the sequence which begins with the NBP and earlier pottery. Nothing else of significance is known except the fact that the ancient Gujarat ports must have been major players in the Roman trade. There are only four reported major early historic sites in eastern Rajasthan: Rairh, Sambhar, Nagari and Bairat. Nagari and Bairat were fortified and certainly go back to the third century BC and somewhat earlier.