The outermost site of this civilization is Shortughai in north-eastern Afghanistan, about 25 km away from the Kokcha, a tributary of the Oxus. The results of excavations at this 2.5 ha site have clearly established its Indus civilization identity, and show that ‘no one lived on this part of the plain before the Harappans’. At the same time, the settlement shows a close adaptation to the landscape, down to the establishment of a canal system which brought water to the site from the Kokcha using a skilful lay-out and the local topographical features. On the other hand, flax seeds found in a ploughed-up field outside the irrigated tract suggest dry-farming too. It is inconceivable that as a Harappan site Shortughai existed in isolation; the possibility of other Harappan sites being located in this area is indeed strong, especially in view of the miscellaneous evidence of Harappan contact (which we shall consider later) with the Bactrian area of north Afghanistan, a little to the west of the area where Shortughai is located. Shortughai could not have been reached without passing through south Afghanistan and crossing the Hindukush, but no positive sign of Harappan contact has yet been obtained from this sector of Afghanistan. Obviously, the location of Shortughai is still problematic and nothing can be said about it beyond a general inference that it was probably located where it is because of the access it provided to lapis lazuli of Badakhshan and possibly tin and other resources of central Asia. This site also brings to the fore the significance of the Oxus-Indus interaction zone to which we have drawn attention in the context of the Baluchistan archaeological assemblages.
What is interesting about the location of Harappan sites in Baluchistan is that they are localized in the areas which witnessed considerable village-growth in the earlier periods. Three clear factors have played roles in their distribution in this region. On the Makran coast, the two major reported sites—Sutkagendor (excavated) on the Dasht river and Sotka-koh (unexcavated) on the Shadi Kaur—have been credited with being ports of the Indus civilization’s maritime links with the Gulf region and Mesopotamia. An equally strong, if not stronger, possibility is that they served as the entry points to the interior of this section of Baluchistan through the otherwise difficult Makran coastal range. Beyond the coast, in the Turbat oasis, the Harappans settled at Miri Qalat and perhaps in the Panjgur oasis as well. The agricultural fertility of these oases and their location on routes made them the foci of Harappan and earlier settlements. Proceeding east, the Harappans settled at Balakot in the Khurkera alluvial plain at the edge of the Sonmiani Bay, both for agriculture and access to the marine resources. In the interior of the south Baluchistan hills there do not seem to have been any Harappan settlements; it is more or less the evidence of Harappan contact that one has to look for at such sites as Kulli, Mehi and others, but again, the Harappans could not have ignored the network of both internal and external routes passing through these areas and their importance as a source of metals. The same situation prevails in the Kalat plateau and the Quetta valley, although the Bolan route to Afghanistan must have been very significant to the Harappans, because the major site of Nausharo (excavated) is in the area of entrance to this pass and was dependent both on this significance and on the general agricultural possibility of the Mehrgarh area. In the north-east, the Harappan level of Dabarkot (unexcavated) was possibly due more to its importance on the Afghan route through the Gomal valley than to anything else.
Fig. 17. Floor design in mature Harappan at Balakot. An identical design occurs on top of the Bodhi throne (3rd century BC at Bodh Gaya)
Another major site, Pathani Damb, is clearly important because of its position at the Baluchistan mouth of the Mula pass across the Kirthar range. On the whole, the Harappan sites in Baluchistan continue the settlement pattern established in the earlier periods: ‘only in those areas which are still agriculturally viable and lie on arterial routes of the region’. Added to these, one has to bear in mind the locations of lapis lazuli in the Chagai hills and copper and lead in the hills between Nal and Las Bela.
The earlier occupational sequence in the Gomal valley was topped by a Harappan occupation at Gumla, and there is at least one more reported Harappan site in this area. It is not difficult to guess that they were there both for the agricultural potential of this area at the edge of the Indus alluvium and for the Gomal route to Afghanistan. No Harappan site is, however, reported from Bannu and the Potwar plateau.
The configuration of the Indus-Hakra plain, as drawn by Louis Flam on the basis of aerial photo-composites and ground surveys, is crucial to the understanding of the locations of Indus civilization sites in Sind. His basic map shows a large part of Sind lying between the channels of the Indus on the west and the Hakra on the east, with a large part of the present delta below modern Hyderabad being possibly a part of a large inlet of the Arabian Sea. The Rann of Kutch was a part of this inlet, and thus, the present-day Kutch would have been an island or a group of islands in a tidal sea. Similarly, it is possible that Saurashtra also was cut off from the mainland Gujarat. What is interesting about the map produced by Flam is that it shows both Shikarpur and Mehar (with Mohenjodaro in between) on the left bank of the Indus, whereas these days they are located on the right bank of the river. Although 20 Harappan sites have been reported by him in the lower Indus basin, including about 13 sites between the Indus and the Hakra, it is difficult to be certain about the nature of distribution of these sites east of the Indus. West of the modern Indus the sites, with the exception of a few like Mohenjodaro, fall mostly in the Kirthar foothill zone and Kohistan where their locations can be understood in relation to the landscape.
Mohenjodaro is located in the agriculturally richest part of Sind, i.e. the fertile and rice-growing plains of Larkana. Shikarpur, the famous entrepôt of commerce located almost at the mouth of the Kachhi plain and thus of the caravan trade through the Bolan pass, lies in an area which is only a little upstream from Mohenjodaro, and it is just possible that in protohistory the Bolan trade came down to the plains and the river at Mohenjodaro. From the Mohenjodaro side a shorter approach to the Bolan pass across the Kalat plateau is through the Mula pass, and we have already noticed the location of a major site (Pathani Damb) on the other side of this pass.
Along the western border of Larkana is the Kirthar range (of its total length of 240 km, 90 km are in Larkana). In the Kirthar foothill zone agriculture is limited and ‘depends chiefly on irrigation provided by damming the torrent-beds when there is a rush of water from the hills during the rains.’ Some of these water-courses are spring-fed and thus perennial. Their water is used for irrigation, ‘being carried off wherever necessary by means of channels.’ There are also a few lake depressions in the Larkana area, the most famous being the Manchhar lake. In the map prepared by N.G. Majumdar on his explorations in Sind there is a cluster of sites around this lake which he identifies with the fishing settlements of the Harappan period. Another cluster of sites lies in Kohistan, further to the south. Agriculture depends here on spring water and moderate rainfall. Majumdar also mentions some local routes passing through this section and apparently, the location of sites here was dependent not merely on easy access to spring water but also on their position along these routes. Towards Karachi he reported two sites—Amilano and Orangi—and more recently, another site—Ahladino—has been noted in the Karachi area. To the east of the Indus, Majumdar’s major site is Chanhudaro, almost due west of the site of Amri, with a ferry crossing across the Indus at this point.
More data have now begun to emerge from upper Sind, especially from the area of the Sukkur-Rorhi hills where huts of Harappan workmen who quarried local flints have been found in addition to a number of sites along the present Nara canal which was once linked to the course of the Hakra. A major site here is Lakhianjodaro. A heavy concentration of sites has been reported between the Indus and the Hakra bed, possibly along the spill channels joining these two systems.
The largest cluster of 174 Indus civilization sites is around and to the west of modern Derawar fort on the Hakra. In the west they extend up to Rahimyar Khan but this may also mean that the sites further west could have been buried under drift sands or simply destroyed. According to the explorer Mughal’s estimate, 79 of these sites are ‘purely industrial sites, meant exclusively for the large scale production of fired bricks, pottery, and items of daily use, including perhaps metallurgical activities such as the melting if not the smelting of copper or bronze’. In addition, 33 settlement sites possess kilns, which suggests specialized craft activities as a part of the habitational complexes. There are 50 purely settlement sites and the rest are so-called camp-sites, possibly indicating the existence of nomads. The following size categories are known: up to 5 ha: 44 sites; 5.1–10 ha: 20 sites; 10.1–20 ha: 8 sites and over 80 ha: 1 site. The largest site of this cluster, Ganweriwala Ther, has two close-set mounds (503 × 533 m and 488 × 290 m; maximum height 8 m) covering an area of 81.5 ha.
In Pakistani Punjab there are sites in the Bari Doab (between the Sutlej and the Ravi), mostly along an old bed of the Beas. Harappa, located on an old bed of the Ravi, is in this sector. Sites are found sporadically even beyond the Ravi but are focused around Sheikhpura west of Lahore. Further west, till the Indus, there is no site in Pakistani Punjab, which is in keeping with the arid conditions of the Sindsagar Doab between the Jhelum and the Indus. There is no site in the Jech Doab (between the Chenab and the Jhelum) either.
To the east of the Sutlej lies the transitional zone between the Indus and Ganga systems which is known as the Indo-Gangetic divide. Its boundaries lie between the Sutlej and the Yamuna and between the Himalayan foothills of Indian Punjab and the course of the Ghaggar (Hakra in Bahawalpur, Pakistan) in Rajasthan. This is a vast area, comprising mainly the Indian states of Punjab, Haryana and Delhi. From the foothills beyond Chandigarh and the Aravalli ridge of Delhi the sweep of alluvium till the Rajasthan desert near Suratgarh is impressive. From the present point of view what stands out clearly in this region is the number of drainage channels which come out of the foothills and once combined in various ways to form the channel of the Ghaggar-Hakra, identified with the Sarasvati of the ancient Indian tradition. All the drainage channels of this region are ephemeral and the reason for the drying up of the Ghaggar-Hakra is an important issue by itself. What has been inferred, primarily on the basis of the palaeo-channels traced through remote-sensing imagery, is that the present-day Sutlej and the Yamuna were once flowing into the Ghaggar-Hakra channel and after they were captured, respectively, by the Indus and the Ganga, the Ghaggar-Hakra system dried up, although retaining some water in the upper reaches during the monsoon. This extinct river system was once alive and its impact on the human settlement can be imagined. One has to visualize, during the time of the Indus civilization, a great swath of alluvium which began at the mouth of the Rann of Kutch, included the Indus-Hakra Doab in Sind, expanded to take in the Punjab plain between the Ravi and the Sutlej, and moved up along the Hakra at the edge of the Great Indian Desert (Thar in India and Cholistan in Bahawalpur, Pakistan) to the Simla foothills and Delhi and beyond. It must also be remembered that climatically this entire belt falls in the arid/semi-arid zone.
It may be interesting to know at this point that the whole of Sind, Baluchistan and a large part of the Hakra system and Punjab plain enjoy only 80 to 160 cm of annual rainfall, and in this large area there is a central sector covering parts of Baluchistan and the central Indus valley, which receives rainfall of only up to 80 cm. The next two annual rainfall belts of 160–240 cm and 240–320 cm are in the next section up to Lahore and Kutch. The rest of the area, including the Punjab foothills, Delhi, Saurashtra, etc. experiences an annual rainfall beyond 320 cm, going up to 640 cm in certain sections. The much-debated issue of the ‘Indus’ climate has been undecided regarding the suggestion of more rainfall in the protohistoric past. If anything, the analysis of calcium carbonate nodules from Harappa suggests a dry climate. According to a report in the Eastern Anthropologist (1992) ‘the carbon isotope ratios of pedogenic carbonate in inner portions of nodules forming at Harappa reflect an arid climate with a very low soil respiration rate.’ There is also doubt about the date of desiccation of the Ghaggar-Hakra system. Micromorphological studies of soil in Haryana have led to the inference that right in the first phase of human occupation of the region, the Ghaggar was a dying river, flowing in shallow braided channels through a 6–8 km wide bed and receiving its water mostly during the monsoon.
A site count published in 1984 showed about 138 classic Indus civilization sites in Uttar Pradesh, Haryana, Punjab and Rajasthan. The Uttar Pradesh sites (Saharanpur and Meerut districts) are all from the Ganga-Yamuna Doab. In the Indo-Gangetic divide zone of Indian Punjab, Haryana and Rajasthan the relevant site-count is 107. The largest site—Rakhigarhi in Hissar district—is said to be as large as Harappa. Excavations have just begun at Rakhigarhi, but the extent of the other major excavated sites of the region—Kalibangan (c. 12 ha or more), Banawali (c. 5 ha or more) and Mitathal (c. 7 ha) is much less. In Uttar Pradesh, the major excavated site is Hulas (c. 5.5 ha) in Saharanpur district. A serious problem regarding site distribution in this region is the apparent cluster of very large sites (one as large as 225 ha, but there are 4 others with 100 and more hectares) in the Mansa area of Bhatinda district, Punjab. These sites are unexcavated, briefly reported, and no explanation regarding their size is forthcoming.
The classic Harappan site count in Gujarat in 1984 was 101:30 in Kutch; 12 in mainland Gujarat, and 59 in Saurashtra. The maps show their distribution along the dendritic river systems of the Kutch and Kathiawar peninsulas and the estuarine regions of the mainland coast. Dholavira, which is more than 100 ha (approximately half of it covered by walled settlement), is the largest site in Kutch, but the other Gujarat sites are much less in extent. Lothal, Surkotada, Padri, Kuntasi, Rangpur and Rojdi are among the major excavated sites of this region, which, because of its coast and raw materials (cf. semi-precious stones, copper and other minerals) must have been important in the Harappan scheme.
Apart from its basic geography and site localization, perhaps the most important characteristic of the distribution of classic Harappan sites is their adaptation to the diverse settings from Baluchistan to the upper section of the Doab and from Manda in Jammu (as an extension of the Punjab sites) to the estuarine area of mainland Gujarat. In this task of regional adaptations they were preceded in all the areas by the cultural level which is now almost universally called ‘early Harappan’. At the same time, the distributional picture which has emerged is a flat picture. Does it mean that the classic Indus civilization appeared more or less simultaneously all over its basic distributional zone?
It is improbable that it did so. On the basis of the sheer density of site distribution, both in the classic Harappan and earlier periods, the core area of the growth of the Indus civilization seems to be the explored Cholistan section of the Hakra plain: Hakra ware sites: 99; early Harappan sites: 40; mature or classic Harappan sites: 174. Significantly, they all lay in an area which measures, along the Hakra course, only about 200 km and not more than 50 km at any point across the breadth. Once it had developed in this apparently limited area, it was not at all difficult for this civilization to come down, along the Hakra, to Kutch first, and then to Saurashtra. It should also have spread quickly to the Hakra-Indus Doab including the Mohenjodaro area, as Mughal’s map (Mughal, 1997) suggests. Similarly, central Indus valley up to Harappa between the Ravi and the Sutlej came in for occupation. The Indo-Gangetic divide too was possibly occupied during this phase. The Baluchistan settlements or the settlement of Shortughai were probably parts of the thrust towards the Mohenjodaro area, while the upper section of the Ganga-Yamuna Doab was possibly settled later. All these areas, except the Doab section, have clear archaeological evidence of earlier settlements; in the Pakistani Punjab plains and at least in a part of the Indo-Gangetic divide the first phase of these settlements is marked by Hakra ware. In the rest of the plains and the Gujarat region the substratum is not earlier than early Harappan. In Baluchistan, of course, there is a much earlier beginning of settled life. In the Doab there is no documentation before the classic Harappans moved in, and in Gujarat the microlith-using hunter-gatherers lay in the background. On the whole, the distribution density is interesting: concentration and a long time-depth in a particular area, and a reasonably dispersed character with a lesser time-depth of earlier occupation elsewhere. One would also infer that the process of geographical dissemination was a more or less continuous phenomenon over a long span of time. The Indus-Hakra alluvial system alone offered them the opportunity of a tremendous and uninterrupted expansion, in the course of which the settlements could come up as defined and dispersed units, at least as long as the classic form of this civilization lasted.