distribution and features of the Indus Civilization

II.2. Distribution and Features of the Harappan Settlements

Editorial Note

[The present essay offers a straightforward account of the distribution and the basic features of Harappan settlements in a given geographical area. It does not attempt any phasewise discussion mainly because the phasewise reporting of sites is not clear everywhere. What it attempts to give is only a collective picture of the settlements, outlining the aspects which emphatically underline their Harappan character. One of the hypotheses which we find difficult to accept is that there is a chronological slope from the west to the east in the distribution of Harappan sites which in the east wash up to the slopes of the Siwaliks, both in the two Punjabs and western Uttar Pradesh. According to our understanding, the Harappan settlements were securely ensconced in the upper part of the GangaYamuna Doab by the middle of the third millennium BC. Going by the date of $c .2200$ вс which was derived not from the lowest level of occupation at Alamgirpur, the date of 2500 BC for the lowest occupational level of Alamgirpur seems to be a reasonable estimate. Equally intriguing is the notion that there were regional traditions of Harappan development in different areas. For instance, the notion of ‘Bara culture’ (thus named after a site near Rupar, a full-blown Harappan site in the foothill belt of Indian Punjab) seems to have crept in the analysis of the Harappan assemblage at the relevant sites of western Uttar Pradesh and Indian Punjab. We on our part find the whole concept of ‘Bara culture’ nebulous and unwarranted. The display of the Bara material in the Archaeological Survey of India museum at Rupar makes us feel that it marks nothing more than a phase of the Late Harappans. In an area as large as the distribution of the Harappan sites, it is not surprising that there will be some variations in regional pottery styles but that does not take away from the overall homogeneity of the Harappan tradition. Besides, the regional differences still remain to be detailed item by item; mere references to the vernacular geographical terms such as Sorath,

Anartta, etc. without spelling out exactly what their characteristics are will not be academically fruitful or revealing.

Perhaps the most striking aspect of the Harappan settlements is the ability to put their concept of planning in well-executed forms with proper well-demarcated areas. There apparently was a background of what came to be considered later on as Silpasastra texts. The fact that there was a series of water reservoirs at Dholavira, some of them hewn partly out of bedrocks, is reminiscent of a later day ideal of a city with a number of water bodies for bringing about an air of peace and coolness in it. From whichever way we view it, the Harappan settlements encapsulate a lot of the Harappan ideal of the world. The present essay is nothing more than an outline of their general picture.]

h

I. NORTHEAST AFGHANISTAN

This is the only area outside the Harappan orbit where a proper Harappan settlement has been found. The name of the site is Shortughai (Francfort 1984, 1989) and it was excavated in the 1970s by French archaeologists. It lies in the Ai Khanum Plain which is a 27 km long and less than 9 km wide valley in the area of the confluence of the Kokcha River with the Oxus or Amu Dariya. The Kokcha rises in the hills of Badakhshan, famous for its lapis lazuli mining. Badakhshan in the northeastern part of Afghanistan is the major source of lapis lazuli, a deep blue and translucent stone famous both as a precious stone and a source of aquamarine and blue pigments. Badakhshan was most likely the source of lapis used in Mesopotamia and Egypt where there was a great demand for this stone. The Ai Khanum Plain is famous in archaeology because of the presence of the Hellenistic city of Ai Khanum there. This city was established around 330/305 вс and lasted till at least 145 вс. This city has been excavated on a large scale by French archaeologists. The
evidence suggests that it was closely linked with the cultural developments in the mainland Greece.

The discovery of the Mature Harappan site of Shortughai in this sector of Afghanistan is an important event. This is the only Harappan site found so far outside the subcontinent. The site, archaeologically the earliest settlement site of the Ai Khanum Valley, has been related to a network of canal irrigation which has been attributed to the presence of the Harappans in the valley. There is no sign yet of any earlier tradition of irrigation in the area. The Harappans must be credited with the introduction of irrigation system in the Kokcha Valley. Actual canal beds have been identified and dated on the basis of the date of pottery found in them. Local topographical features were used to draw water from the Kokcha. The climate of the time has been inferred to be similar to the presentday climate of the area. Agriculture was necessarily dependent on irrigation. A ploughed field covered with flax seeds has been excavated here, and this discovery suggests that dry farming was prevalent in the areas which could not be covered by irrigation.

Shortughai is the earliest of the seven Bronze Age sites of the plain and about 25 km away from the Kokcha and about 5 km away from the Oxus. The site is 2.5 ha in extent and shows two mounds, A and B. In Period I only a part of Mound A was occupied. The thickness of the occupational deposit of Period I is only 50 cm , but it shows a broad range of Mature Harappan objects and other features: plain and painted pottery of typical Harappan forms and designs; terracotta cakes, toy-carts, spoons and figurines; a discoidal mirror of copper and fragments of other copper objects including clay crucibles with copper inside; gold and lead fragments including a discoid gold bead; evidence of local

Fig. 1. Important Harappan sites mentioned in the text.
processing of lapis lazuli and carnelian; objects of lapis lazuli, agate, turquoise, carnelian, steatite and sea-shells (bangles of xancus pyrum); an inscribed Harappan seal depicting a rhino and inscribed potsherds; and finally, well laid-out houses made of mud bricks of the standard Harappan size (1:2:4 ratio). It has been emphasised that “all the above recorded antiquities of Shortughai exhibit purely Harappan shapes, sizes, proportions and decorations.” Nothing in this range of objects can be attributed to any non-Harappan culture. The
excavated plant remains are barley, wheat, panicum millet, lentil, pea, almond, pistachio, grapes and linseed and they have all been parts of the Mature Harappan complex of the subcontinent. So, even from this point of view, Shortughai remains an indisputably Mature Harappan site.

Shortughai seems to be very much a lone Mature Harappan outpost in the northeastern corner of Afghanistan. Looked at closely, the site lies almost at the border of Tajikistan. Why there should be a Harappan site here is a historical

problem. Its immediate association is with the mining and trade of lapis lazuli in its immediately adjacent area which also possesses some sources of tin. The area also has additional geo-political importance as a convenient position to maintain close links with the wide region between Turkmenistan in the west and Chinese Turkestan or Xingxiang in the west. Jade occurs in the Harappan context and the likely source of that jade was Chinese Turkestan. There is also evidence of contact between the Harappans and the sites of Turkmenistan, especially Namazga Tepe. So our inference that the site has a geopolitical importance is not an improbable one.

The reports of the Harappan materials such as etched carnelian beads and steatite seals in the antiquities market of Afghanistan suggest that although Shortughai is currently a lone Harappan outpost in Afghanistan, there may be more discoveries of this kind in the future. The radiocarbon dates of the site from Period I are around the middle of the third millennium BC , but some of them date from the first half of the third millennium BC .

The route to Shortughai from south Afghanistan is from the Begram Plain via Charikar and the Khawak Pass, the latter at an elevation of 4000 m . To reach Shortughai the Harappans were crossing the high passes of the Hindukush. They were possibly reaching the Begram Valley as a staging post for Shortughai and we suggest that the Harappan approach to the Begram Plain was possibly through the Ghazni area of south Afghanistan, Ghazni being easily approachable through the Gomal Valley. There are Harappan sites at the mouth of the Gomal Valley. Dabarkot located in this section is a major site. Another possible reason for the location of Shortughai was that it acted as an entrepot of horses which could and generally
did easily come from Tajikistan, into early historic India.

II. BALUCHISTAN

Many sites in Baluchistan (Stein 1929, 1931, Franke-Vogt website, Fairservis 1971, Dales 1979, Besenval 1994, 1999) reveal signs of contact with the Harappan Civilisation, but the sites with distinct Harappan occupational levels are not many. In northeast Baluchistan the major Harappan site is Dabarkot. This shows thick Harappan deposits high up in the section of a mound which is about 376 m in diameter. The site is a part of the Gomal system which marks the route toward the Ghazni-Kandahar area of south Afghanistan.

The Kachi Plain has two major sites, Judeirjodaro in the southern section of the plain and Nausharo in its northern edge near Mehrgarh. These sites lead toward the mouth of the Bolan

Fig. 2. Nausharo site plan (after Jarrige 1989).

Pass, but as the Kachi Plain is agriculturally prosperous and has a good reputation for local bullocks, it is possible that Judeirjodaro and Nausharo might have been agriculturally important settlements in their own right. Judeirjodaro shows a cluster of mounds covering an area of $549 \mathrm{~m} \times 364 \mathrm{~m}$. The rain gullies of the site may indicate the lay-out of streets. It has a main mound and five smaller mounds near it. In its excavated form Nausharo has revealed a Mature Harappan settlement with areas separated by parallel streets, huge mud brick platforms, houses with bathrooms connected by terracotta pipes with the soakage jars outside, and house foundations formed by wooden beams set in pebbles and clay. There is also evidence of possibly a large tank or canal with a spillway made of burnt bricks. The hills on the western side of the Kachi Plain contain the site of Pathani Damb at the mouth of the Mula pass on the Baluchi side. It covers a large area and a high central mound. A report of the 1960s described it as being equal to Mohenjodaro in size and importance. No further work seems to have taken place at this site.

The following Baluchi areas are important for the distribution of Harappan sites: the Las Bela Plain and the adjacent hill valleys at the border of Sind and Baluchistan, the Makran Coast, and the Kej Valley to the north of the

Fig. 3. Large painted jars from Nausharo.

Fig. 4. Ceramic from Nausharo ID showing transition from Early to Mature Harappan phases (after Jarrige 1989).

Makran coastal range. Balakot is the main Harappan site of the Las Bela Plain, its precise location being the Khurkera alluvial plain near the Sonmiani Bay. This site, which measures about 3 ha in extent, has a higher western mound and a lower eastern one. There is some evidence of lanes and mud brick houses. The use of burnt bricks was observed only in some drains. One room has its floor paved with square tiles bearing impressed intersecting circle designs. A courtyard was found lime-plastered and it had a circular depression with the remains of a wooden column at the centre. The threshold of the houses was made of wood and a pottery tub was found

in the room which had its floor decorated with intersecting circles. A hearth, a buried storage jar and a drain formed by a broken pot were also found in this room. The ordinary room size at Balakot was $2.20 \mathrm{~m} \times 3.20 \mathrm{~m}$. One large kiln was found associated with a few smaller ones. These kilns were used possibly for baking terracotta animal figurines. There is no evidence of a surrounding wall or fortification at Balakot. On the western side of the Las Bela Plain there is a smaller Harappan site called Khairiakot.

The Kanrach Valley is in the hilly section adjoining the Las Bela Plain on the east, with the seasonal stream of the Kanrach flowing through it. The valley is framed by the Mor (1400 m ) and the Chapar ( 1500 m ) ranges. Bakkar Buthi, a small Harappan site located on a terraced

Fig. 5. Ceramic from Nausharo ID showing transition from Early to Mature Harappan phases.
hill above the Kharari (a tributary rivulet of the Kanrach) and overlooking its valley, comprises a fortified southern mound and about two houses and open spaces related to chert production outside it. Much of the pottery is identical with that found at the classical Harappan sites but there is also an element of local production. A huge stone-built dam blocked the Kharari before it entered the Kanrach Valley and was the easternmost in a series of three dams. The site has been dated around 2400 вс. More to the southeast, the Hab Valley provides access from Sind to the interior of Baluchistan. The area seems to be dominated by ‘Kulli’ sites, many

Fig. 6. Ceramic from Nausharo IC showing transition from Early to Mature Harappan phases.

with well-built dams nearby, but there is no specific mention of a Harappan site in this sector.

The main Harappan occupation in an oasis which falls on the main highway coming from eastern Iran is represented by Period IV of the site of Miri Qalat. The Turbat oasis is to the north of the Makran coastal range. The excavated remains comprise mud/mud brick houses and miscellaneous artefacts such as an ivory comb and a steatite seal. The Harappan material occurs alongside the ‘Kulli’ remains, and another interesting feature of the site is the profuse occurrence of fishbones of presumably marine origin. Prahag, which is situated on a sand dune in the Pasni Plain near the coast, is also a Harappan site and has been interpreted as a place where large sea fishes were processed.

Sutkagendor and Sotka Koh are located in the Dasht Kaur and Shadi Kaur valleys respectively and known as two well-known coastal Harappan settlements of Makran. The location of Sutkagendor is about 50 km away from the sea. It was once suggested that the settlement was once positioned on the sea-shore and that it was a port in the sea traffic between Sind and the Gulf region. However, no positive evidence has yet emerged in this direction, and, according to us, a better and more convincing hypothesis is that it is located on the route which led from the coast to the interior through the Dasht Valley. The site (c. 180 m x c. 103 m ) was converted into a fortified enclosure by building thick ( 7.5 m wide at the base on the eastern side) walls made of stone blocks set in mud to join two natural ridges. Inside the fortified enclosure there was about 3 m thick Harappan occupation deposit. There are traces of gateways and bastions along the fortification walls. The general occupational area outside the fortified enclosure was flimsy. Sotka Koh of the Shadi Kaur Valley resembles Sutkagendor in its lay-
out but one does not yet know if this was a completely enclosed fortification. Like Sutkagendor, this could also mark a route from the Makran coast to the interior. This route went through the Shadi Kaur Valley (Mockler 1877, Stein 1931: 60-71, Dales 1962).

SIND
The geography of the modern province was in all probability different in the ancient context. In the light of the argument given by L.Flam on the basis of historical data, aerial photographs and ground surveys, the Indus was then following a different course, joining the Ghaggar-Hakra flow at Naukot and then flowing into the sea near the modern Rann of Kutch as the combined stream. It has been further presumed that the coastline and the mouth of the Indus River were then as far north as the south of modern Hyderabad. One of the implications of this idea is that the modern territory of Kutch, instead of being a peninsula, was then possibly an island and the same might have been true of the Saurashtra peninsula as well. Besides the delta, two areas of Sind were important in the locations of the Harappan sites: the Kirthar piedmont zone, which is dissected by the torrents coming down this rampart-like hill system between Sind and Baluchistan, and the zone of Sind Kohistan which is an attractive area of occupation because of its open valleys, low hills and thermal springs. In the Kirthar piedmont zone agriculture is dependent on check-dams which are built to check the flow of the water torrents coming down the hills. Some of these torrents are also perennial, being fed by springs. None of the Harappan sites reported in this area is a major one.

The western edge of Larkana, where Mohenjodaro is situated, possesses a few marshy depressions. The most important of these

depressions is Lake Manchar where N.G. Majumdar in his Explorations in Sind in 1934 reported a cluster of Harappan sites interpreting them as fishing settlements. In the Kohistan section, agriculture flourishes on the basis of the water that is available from springs and annual rainfall, the latter only a moderate amount. There are also local routes between Sind and Baluchistan in this area. Its Harappan sites are related to such routes and agriculturally suitable spots. The sites of Amilani, Orangi and Ahladino are oriented towards Karachi. L. Flam’s map shows only 20 Harappan sites in this sector, out of which 13 are located between the Indus and the Hakra. According to Flam, the sites in the lower Indus Basin increase from 3 in the AmriNal phase to about 20 in the Harappan context. Northern Sind has more sites and some of them are working men’s quarters in the Sukkur-Rorhi Hills where the chert for the famous Indus ‘Sukkur-Rorhi blades’ was quarried. The Lakhianjodaro is a major site of this sector, and this and the other Harappan sites of the area are found along the spill-channels joining the Indus and the Hakra (for distribution of Mature Harappan sites in Sind, Majumdar 1934; Flam 1986; Biagi and Cremaschi 1990).

Around 1834, Sind was a rich agricultural area and an idea of how agriculture was conducted here comes from a report by J. McMurdo (1834). Sind was known to have two sections, the southern section called Lar and the northern one called Sirra. The lower part of Lar was subject to prolonged flooding by the Indus and produced enormous quantities of rice. However, this section south of Hyderabad was possibly within the area of the sea in the protohistoric period and thus its agricultural prosperity in 1834 is not relevant to the protohistoric context. However, McMurdo’s comments on the agriculture of the north are also interesting:

The more northern districts produce abundant crops of wheat, barley, jowari, mung and other common grain, almost exclusively, however, in many parts by irrigation, or in the moist beds of extensive dbands or lakes which, formed by the annual floods, gradually but quickly evaporate.

Mohenjodaro is in the Larkana area, and Larkana’s agricultural fertility was famous as late as the early 20th century: “perhaps the finest tract in the whole province” which produced three crops a year including wheat and perhaps the finest rice in the province along the riverbanks. The location of Mohenjodaro also gained importance from the fact that like Shikarpur of the same area, it was convenient enough for a smooth access to the Kachi Plain and the Bolan Pass. I would say that the trade coming through the Bolan Pass reached the river-bank at Mohenjodaro in the protohistoric period just as it did at Shikarpur in the 19th century. From this point of view, Mohenjodaro was the protohistoric ancestor of Shikarpur of the 19th century.

Mohenjo-Daro (Mohenjodaro, Marshall 1931, Mackay 1938, Dales 1965, Dales and Kenoyer 1986, Jansen 1987, 1993)

Excavations began at Mohenjodaro in 1922 under R.D. Banerji who had visited the site several years earlier for its Buddhist stupa. The site was visited for the same reason early in the 20th century by D.R. Bhandarkar, another Archaeological Survey of India officer. Between 1922 and 1927, the Mohenjodaro excavations were under the general direction of John Marshall and between 1927 and 1931 the responsibility rested with E. Mackay. The site was subjected to brief excavations in 1950 by Mortimer Wheeler on behalf of Pakistan’s Department of Archaeology and in 1964-65 by G.F. Dales of Pennsylvania University. On the western bank of the Indus, the site is a little away from the riverbank. In 1931 the distance

between the site and the river was reported to be about 5 km whereas in the 1990s it was reported as only 1 km .

More alarming than the river for the safety of the site is the problem of the rising level of the sub-soil water and the consequent problem
of salinity. In 1922 the groundwater level was about 8 m below the surface but in 1981 it fluctuated between 1.52 m in October and 3.66 m in May. The salt present in the sub-soil water can now more easily rise to the surface because of capillary action, and, as the area has a high

Fig. 7. Citadel and Lower Town at Mohenjodaro.

summer temperature and a low rainfall, the surface moisture rapidly evaporates, leaving the salt concentrated on the ground and thus on the excavated remains. The excavated structures are slowly crumbling, and the future survival of Mohenjodaro as a site will depend on the extent to which the modern conservation efforts are successful. The fact that the Indus is moving closer to the site is a problem, but the salinity which affects the excavated bricks and leads to their decomposition is an immediate problem of greater magnitude.

A 200 m wide open space separates the western and the eastern mounds of Mohenjodaro. As this space did not represent an ancient flow of the river, it has to be considered as being originally open and empty. The area of the western mound is about 400 m x 200m, standing on an artificial platform of clay which in turn is enclosed by a 6 m thick mud brick retaining wall. The height of this platform was about 7 m after two enlargements. Further platforms of varying height were constructed on it as the podium of more buildings. The retaining wall was also raised higher and this resulted in the creation of a fortification wall around the mound as a whole. The fortification wall was made complete with bastions and probably gateways and stairs to reach the ground level. There was a platform at the base of the eastern mound too, but the height of this platform was less than that of the western one. In the northern section of the western perimeter of the eastern mound, there was a wall which was found to be more than 9 m thick, and in the southern section on this side there was a “massive construction composed mainly of huge solid mud brick embankments with baked brick retaining walls”. It is probable that there was a retaining wall around the platform on the eastern side too. The original area of the eastern mound is somewhat
uncertain. The present irregularly shaped area is about $820 \mathrm{~m} \times 720 \mathrm{~m}$ wide, but there are traces of occupation, dating possibly from the late phase of the city, for about 2 km to the east of the site. The ancient city in its late phase possibly encompassed an area up to 200 ha. The earlier estimate of its size, i.e. 240 or 250 acres (c. 101 ha) does not seem to be accurate.

The high sub-soil water at the site prevented the excavation trenches from reaching the natural soil. Borings at the edges of the site showed traces of occupation in the form of pottery and other debris much below the level reached by the excavation trenches. The premise of seven successive levels of the site suggests that there are three main phases-Early, Intermediate and Late. However, the earliest occupational level is designated Early III. The assumption is that the first two sub-periods of ‘Early’ (i.e., Early I and Early II) are still undug. The ‘Intermediate’ and the ‘Late’ have 3 sub-phases each, with a further sub-division of the Late I into the Late Ia and the Late Ib. Apparently, a vast amount of clay would have been required to build the platforms at the base of the western and eastern mounds, and it is probable that this clay was dug out of the immediately surrounding land. If so, this digging for clay would have led to the formation of deep surrounding ditches around the platforms. The rubbish thrown into these ditches would ensure that some of the occupational remains would find their way well below the occupational level of the general surface. What this hypothesis offered by M. Jansen is trying to tell us is that there is no unexcavated stratum at Mohenjodaro. Jansen initiated a very detailed architectural survey at Mohenjodaro, but this hypothesis fails to be convincing because the deep cuttings beneath the pavement of the Buddhist stupa on the western mound did not reach the natural soil of

Fig. 8. Aerial view of Mohenjodaro citadel.
the site, thus proving almost conclusively that the regular occupational strata still lie undug below the bottommost level reached by excavations at the site.

The Kushan period Buddhist stupa, the date of which is also strengthened by the find of coins of this period in the vicinity, marks the highest point of the western mound. The stupa stands in its northeastern corner. Assuming that the spot on which the stupa stands has not changed its sacred character since the protohistoric period, it has been assumed that if there was indeed a temple in the ancient city of Mohenjodaro, that must have stood at this site. A large open court with massive southern and western walls has been excavated to the northwest of the stupa and has been interpreted as the space where temple revenues in kind could be brought, assuming, of course, that there was a temple below the stupa. Interpretations of this kind are
difficult to support without specific data. Similarly, on the basis of the supposed existence of a temple on this spot, a multi-roomed large structure with thick outer walls, which has been excavated across the north-south ‘Divinity Street’ (its width varying from 2.05 m to 3.15 m ) to the west of the stupa, has been called ‘Collegiate Building’ and interpreted as “the residence of a very high official, possibly the high priests”. The ‘Main Street’ ( 3.73 m to 4.47 m wide) runs along the western wall of the Collegiate Building and has at its north-western end a series of rooms with a drain running through a corridor. These rooms have been interpreted as ‘ablution places’ for the select priesthood.

The Bronze Age cities of Egypt, Mesopotamia, Crete and elsewhere had each a principal temple and in the case of Mohenjodaro also this has been assumed to be the case. This

Fig. 9. Details of various buildings in the Mohenjodaro citadel area.
assumed temple of Mohenjodaro has been located at the present site of the Kushan period stupa in the northeastern corner of the site and also supposed to lie below it. The terms like ‘Divinity Street’, ‘Collegiate Building’, ‘ablution places for the select priesthood’ are based on this assumption. There is no way archaeology can support the implications of these terms but somehow they impart the feeling of a living city to the silent ruins of Mohenjodaro.

Fig. 10. Mohenjodaro: One of the main streets.

To the west-southwest of the ‘Collegiate Building’ there is the famous Great Bath. This is basically a rectangular pool set in the midst of an open paved courtyard and surrounded by open galleries and rooms. Its western, eastern, northern and southern sides are, $14.52 \mathrm{~m}, 14.5$ $\mathrm{m}, 6.98 \mathrm{~m}$ and 7.12 m long respectively. Steps of brick covered by wood set with bitumen or asphalt go down to the paved floor of the pool from its north and south sides. The pool is 2.43 m deep from the level of the paved courtyard. The northern staircase has a small platform at the base and a small further step. An outlet at the south-eastern corner is connected with a corbelled drain which goes down the western slope of the mound. The pool used to be filled with water drawn from a well in one of the neighbouring rooms. The bricks used in its construction are set on edge with gypsum. The pool was made watertight by the application of bitumen on its outer wall which lay within another enclosing wall. The intervening space between the two walls was filled up with rammed mud. It has been pointed out by Michael Jansen that this ‘bath’ stands on an earlier foundation and was apparently the only free-

Fig. 11. Aerial view: Great Bath, granary and other buildings.

Fig. 12a. View of Great Bath at Mohenjodaro.

standing structure which could be walked around at Mohenjodaro. A possible explanation is that it was a ritual tank just like the ritual tanks of the later Hindu temples.

What has been called a granary, or rather, the podium of a granary, at Mohenjodaro, lies immediately to the southwest of the Great Bath. It is a structural complex of 27 rectangular blocks of mud brick with criss-cross passages between them over an area of 45.72 m (east-west) x 22.86 m (north-south). There is a brick-built platform with an alcove at its eastern end in front of this complex. The outer walls of this complex are massive and battered, and possess vertical chases which are supposed to hold the wooden structures at the top of each of the 27 mud brick blocks. The brick platform in front has been

Fig. 12b. Details of the Mohejodaro Bath.
interpreted as a loading platform for putting grains into the wooden superstructures. The interpretation of this complex as a public granary is not universally accepted.

The buildings in the ‘stupa area’ seem to be mostly public buildings, but to its south, there are blocks of ordinary houses with bathrooms and drains, which are separated by streets and lanes. In the excavation literature of Mohenjodaro, this is called the L area and here a prominent excavated building is an approximately square pillared hall where 20 rectangular piers of bricks divide the space into five east-west corridors. There was possibly an arrangement of raised benches along the corridors. The explanation of this hall has varied from an assembly hall to a market hall. Interestingly enough, large halls with such raised benches can still be seen as market places in modern India.

The eastern mound of Mohenjodaro is about 5.62 times larger than the western one, although in height it is lower than the western mound. They were, however, constructed in both cases by raising an artificial platform supported by massive retaining walls and by creating further platforms on it for individual buildings. There are differences in detail between the two. Whereas the western or the higher ‘citadel’ mound has in its northern SD area (thus named after its excavator Siddiqi) some prominent nonresidential buildings, the lower, larger and prominently residential eastern mound has been divided into three long slices of habitations by three north-south streets at 180 m intervals. The excavated sectors here have been given in the excavation literature the names of their excavators: HR (Hargreaves), DK (Dikshit), VS (Vats) and Moneer. In the better exposed HR area, the ‘First Street’ ( 9.14 m to 10.66 m wide), two smaller but roughly parallel streets, and a

Fig. 13. Structural remains with standing wells at Mohenjodaro.

Fig. 13a. View of the Great Bath and other buildings (courtesy: National Geographic, 2008).

Fig. 14. Narrow lanes between two-storyed houses.
few lanes contain houses whose entrances are primarily from the sidelanes and where rooms are disposed around courtyards. Drains, occasionally with stone or brick covers, are a common feature of all the excavated sectors. Two brick-built enclosures were found connected by a drain in ‘First Street’, and in one of these enclosures there were steps going to the bottom. It appears that the enclosure was subject to periodic clearance. Little deposits of sand found by the side of the drains indicate that they were regularly cleared. The general depth of the drains is $c .23 \mathrm{~cm}$ and their width $45-60$ cm . Brick-built culverts are reported from the edge of the habitation, suggesting that they were intended to discharge the total drainage outside the habitation. In the HR sector, there is a double row of 16 houses, each with a room in front and one or two rooms at the back. Their size and construction contrast sharply with the

Fig. 15. Mojenjo-Daro city plan, HR Area.

Fig. 16. Lower street, HR area.

Fig. 17. Mohejodaro: Lower Town, view from north.

Fig. 18. Bathroom, Mohenjodaro.
substantially built houses of the area and they have been interpreted as shops /servants’ quarters/coolie-lines. Some houses had upper storeys, as is proved by the stairs found in them. The houses were flat-roofed, the roof being made of either mud-covered reed-frames or rammed mud supported by wooden beams. One house in the VS area could be entered straight from the main road and had circular depressions on the floor possibly to keep large jars. The possibility of this being a shop is high. A house in the HR area (numbered House I) has been interpreted as a possible temple mainly because of its approach through a large gateway and the find of two small stone male figures in its

Fig. 19. A corbelled drain at Mohenjodaro.
precincts. In the DK area, some excavated structures were labelled as ‘governor’s palace’, ‘pilgrims’ hostel’ and ‘khan’ or caravanserai. Such speculations are nothing but excavators’ attempts to infuse some character into the vast piles of burnt-brick houses. One house in the Moneer area had 2 m thick walls and was 890 sq m in size.

The construction of Mohenjodaro must have involved organisation and planning on a huge

scale. Michael Jansen has calculated that the two platforms at the base of the western and eastern mounds, which may be assumed to have been 5 m high at the minimum, would have required 4 million cubic metres of clay and sediments in addition to what was necessary for millions of mud bricks. Further platforms were created on the top of these platforms to create the base of at least important buildings such as the Great Bath. That also would have needed extra labour and time. According to Jansen, “the highest raised buildings must have stood more than 20 m above the surrounding plain and were visible from afar.” Whether his calculation that the platform of the western mound alone would have needed 20000 people working for two winter seasons is true or not, that gives an idea of the scale of planning and organisation behind the construction of Mohenjodaro.

Other Sind Sites

Although much smaller than Mohenjodaro, Chanhudaro ( 6.5 ha ), located by the side of an old bed of the Indus, was linked to a gap across the Kirthars about 60 km away. It showed a 5.68 m wide street with two covered drains and a bead-making place by its side. This place had an abundance of finished and unfinished beads and an arrangement to glaze them by heat. The individual houses were built on solid mud brick platforms. The site, identified by N.G. Majumdar and excavated in 1935-36 by E. Mackay, showed 3 phases of Harappan occupation succeeded by the two post-Harappan phases, known successively as Jhukar and Jhangar. At Ali Murad, another major Harappan site in Sind, a “long rampart wall of irregularly dressed stone blocks” has been found. The site of Kot Diji, well known for its earlier occupation, had about 5 m thick deposit of Mature Harappan level showing traces of ‘a well-regulated town-plan’ with
streets and blocks of houses. The much smaller site of Ahladino (less than 1 ha) near Karachi (mound diameter: c. 100 m ) showed a well set in an open court ( $20 \mathrm{~m} \times 8 \mathrm{~m}$ ) with house remains near it. A clay receptacle, shaped like a modern bath-tub and with inter-circle designs at its bottom, was also found near this well (Mackay 1943, Majumdar 1934, Fairservis 1982, Khan 1965).

The Harappan sites of Sind were located in the fertile parts of the province and were also linked to the alignment of the Indus which must have had a number of major ferry points. Another factor was possibly the alignment of the major route from the direction of modern Karachi to the northern sector of Sind. The location of suitable chert resources was the reason why working quarters of the Harappan period have been found in the Sukkur-Rorhi Hills.

Fig. 20. A computer-generated image of houses at Mohenjodaro.

Fig. 21. Site plans of Mohenjodaro and Harappa.

CHOLISTAN

As far as one is aware, no Harappan site has yet been excavated in Cholistan. Our knowledge of the Harappan sites of this area is almost exclusively dependent on the site-list and distribution maps prepared by Rafique Mughal. It appears that after the Hakra ware and Early Harappan periods the distribution of the Mature Harappan sites shifted somewhat to the southeast of the earlier distribution areas. It is possible that the river, i.e. the Ghaggar-Hakra, changed its course during this period and that the Mature Harappan site distribution reflects this change. Ganweriwala is the largest of the 174 Harappan sites of the region, and at 81.5 ha it is about onethird of the size of Mohenjodaro.

Mughal (Cholistan, Mughal 1997) has made a distinction between the residential and industrial sites of Cholistan, trying to impart
functional attributes to them. Of all the 174 sites, ‘purely industrial’ sites are supposed to be 79. There are 33 settlement sites with potters’ kilns and there are 50 purely settlement sites. The remaining 12 sites are supposed to be camp sites, i.e. sites without occupation of a permanent character. Mughal infers that the purely industrial sites were used exclusively for the large-scale production of burnt bricks, pottery, items of daily use and possibly copper implements. The sites with kilns suggest the manufacture of pottery and similar other items at otherwise normal habitation sites. The camp sites may, or may not, suggest nomadic settlements. The sites which measured in extent up to 5 ha were 44 in number. The size of 20 sites varied between 5.1 ha and 10 ha. The size of 8 sites varied between 10.1 ha and 20 ha , and only one site measured over 80 ha in extent.

Fig. 22a. Blocked corbelled-arch drain, Mohenjodaro.

PAKISTANIPUNJAB

For an idea of the distribution of Harappan sites in Pakistani Punjab (For distribution of sites in Pakistani Punjab, Mughal et al. 1996; for the comment on climate, Dales and Kenoyer 1992) one has to depend on the distribution maps published by the Pakistan Government’s Department of Archaeology. There are sites along an old bed of the Beas in the Bari Doab between the Sutlej and the Ravi. Harappa is located on an old bed of the Ravi in this section. Except a few sites near Sheikhpura near Lahore there is no site beyond the Ravi. The Sindsagar Doab between the Jhelam and the Indus does not seem to have any site and the same is true of the Jech Doab between the Chenab and the Jhelam. There is no site in the Sindsagar Doab between the Jhelum and the Indus and the Jech Doab between the Chenab and the Jhelum. It

Fig. 22b. Corbelled-arch drain at Harappa.
appears that the eastern segment of Pakistani Punjab contains most of its sites and seems to reflect the traditional occupational pattern of the Punjab Doabs before the modern irrigation canals. This seems to be yet another indication of the fact that the climate has not changed since the Harappan times. It has further been observed that " 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". These nodules were collected in the excavations. The gazetteer of the Montgomery district, where Harappa was located in British India, thought that the region was a ‘howling wilderness’, but plenty of agricultural land was available around the site

which measures about 150 ha in extent and has been much destroyed by the brick-robbing which took place during the construction of the Lahore-Multan branch of the Indian railways.

Harappa

The western mound at Harappa is called AB in archaeological literature. The eastern mound is called E while there is a mound F between the AB and the bed of the Ravi. There are some burial sites to the south of the AB. Harappa was noticed by several travellers, including Alexander Cunningham, before the first excavations took place here in 1920-21 under D.R. Sahni of the Archaeological Survey of India. The major excavations took place under M.S. Vats from 1933 onward. In 1946, there was only
a season’s excavation under M. Wheeler. All these excavations were confined to the mounds $\mathrm{AB}, \mathrm{F}$ and the southern burial areas. The work undertaken by the Pakistan Department of Archaeology after 1947 seems to have been limited to the burial areas. The site was given over to an American team in 1986. They continued their work till the recent political disturbances apparently stopped it.

The fortification wall around mound AB is a prominent feature of the site. On the western side this curves inwardly for some length and is looked over by two salients, the one in the northwest projecting 7.01 m and the other in the southeast projecting 4.57 m . The base of the fortification wall, which was strengthened by bastions or salients, was formed by a mud and

Fig. 23. Granary at Harappa.

mud brick rampart which was 3.04 m high. The fortification wall itself was battered both inside and outside, and on the outside it was further strengthened by a burnt-brick revetment. The width of this wall varies from 12.03 m in one excavation cutting to 15.54 m in another. Traces of a ramp or stairs have been found where the defence wall inwardly curves on the west, marking possibly an entrance to the fortification. The wall to the north of this entrance was fronted by terraces which might have been used for ceremonial processions.

The eastern mound or mound E of Harappa was too destroyed to reveal any coherent pattern of planning, but a 10.5 m high platform of mud and mud bricks was found integral to the rampart in this section, thus suggesting that as in Mohenjodaro, the buildings of Harappa also stood on artificial platforms. What has been presumed to be a granary on the model of the Roman period granaries excavated in Britain has been traced on the riverbank in the area of the F mound. It has two blocks, east and west, each being $45.72 \mathrm{~m} \times 17 \mathrm{~m}$ with a central 7.01 m wide and open aisle between them. Each block had 5 halls ( $15.77 \mathrm{~m} \times 5.33 \mathrm{~m}$ ) divided by 5 corridors. These halls had timber floors which rested on three sleeper-like walls with empty space beneath. This empty space might have been intended for a better circulation of air below the floor of the complex. The three sleepers had extra width at their two ends, which implied that they carried the weight of a heavy superstructure. Further, the entire southern side and parts of the eastern and western sides stood on a battered retaining wall. It is possible that the entrance to this structural complex was from the north, i.e. the river-side. This was certainly not a residential building.

To the south of this complex lies a series of circular (c. 3.5 m wide) walls which were made
by setting burnt bricks on edge to make 4 rings with a hollow at the centre. Twenty of these platforms have been found in excavations. They are all equidistant ( 6.09 m to 6.04 m ) from one another. The hollow of one platform yielded burnt wheat and husked barley, and the general resemblance of these platforms with the grainpounding platforms of the modern village in India is striking. However, two recently excavated platforms were found inside rooms and as the platforms of this kind are usually set in the open in the courtyards, this casts some doubt on their identification as grain-pounding platforms.

Fourteen small houses, each measuring $17.06 \mathrm{~m} \times 7.31 \mathrm{~m}$ within a general enclosure wall, lie along three east-west lanes intersected by 6 narrow alleys. They are accessed by an obliquely set passage and two of its inner rooms are partially brick-paved. At a higher level near these houses there are 16 pear-shaped furnaces, with a crucible for smelting copper being found in one. There is no assurance that the small houses and the furnaces were contemporary.

Two main burial areas, Mature Harappan R37 and Late Harappan Cemetery H, are both located to the south of the AB mound.

There are traces of small streets, drains and soak-pits in the northwestern section of mound E. There is also a series of mud brick platforms with burnt-brick retaining walls in this section. A major north-south road on the southern slope of the mound leads to a mud brick wall ( 5.4 m to 6.5 m wide) which has been traced for about 73 m . This wall is pierced by a 2.6 m wide gateway. There are traces of a burnt-brick revetment wall on the southern side of this gateway. It is important to note that the southern side of mound E was protected by a freestanding wall with a gateway.

RAJASTHAN, INDIAN PUNJAB, HARYANA AND UTTAR PRADESH

In western Uttar Pradesh, Harappan sites have been reported roughly between Delhi and the Siwaliks, i.e. the area constituting the upper part of the Ganga-Yamuna Doab. Basically, all these sites possess the painted pottery tradition of Harappan affinity. When excavated, the sites of Alamgirpur and Hulas, one in Meerat and another in Saharanpur, have yielded Mature Harappan levels, but by and large, the idea is that these sites belong to the Late Harappan phase. However, two sites of this area, Mandu and

Sinoli, may force us to revise our opinion of the character of Late Harappan sites. The wealth of gold and beads of semiprecious stones from Mandi, which otherwise is an inconspicuous Late Harappan site in Muzaffarnagar, demonstrates that the Late Harappan phase was hardly a phase of decadence in trade and wealth. The burial complex which has been excavated at Sinoli near Baghpat shows that the associated Harappan settlement site, although not yet identified, was unlikely to have been an insignificant one. The burial ground itself was well organised and contained burials full of

Fig. 24. Citadel and lower city at Kalibangan.

quality burial pottery and a rich assemblage of grave goods. What is impressive is the broad sweep of these sites right up to the Siwaliks near Haridwar. My inference is that sites like Lal Qila in Bulandshahr and a site near Aligarh, both of which yielded painted pottery, are likely to belong to the Harappan tradition as well. These sites, especially Lal Qila, are supposed to belong to the Ochre Coloured Pottery complex, but as a firm relationship has been established between the OCP, ‘Copper Hoards’ and Harappans in the Doab, one would not be surprised if the painted pottery sites of Bulandshahr and Aligarh are included in the list of sites of Harappan affiliation in the Doab.

Harappan archaeology of Haryana is dominated by the problem of the Saraswati and

Fig. 25b. Structural remains along a street, Kalibangan.

Fig. 25a. Main street at Kalibangan.

Fig. 26. Seven fire altars at Kalibangan.
the Drishadvati rivers. At the simplest, the Ghaggar and the Saraswati in their upper courses are separate rivers. The Saraswati is joined by the Ghaggar near the Haryana-Rajasthan border. All the branches of the extinct river systems of Haryana cannot be satisfactorily plotted, dated and correlated to the distribution of archaeological sites. It is also possible that a good number of sites of this area are linked to an effective canal system involving the two main rivers and their tributaries. Although there are many geological opinions about the GhaggarSaraswati system and the causes of its drying up, it is virtually impossible to come to any satisfactory academic judgement about them.

Kotla Nihang Khan in the upper Sutlej Valley, which is virtually in the Siwaliks, was reported in Vats’ excavation report on Harappa in 1940. In 1951-52, A. Ghosh published the distribution of sites in the Saraswati-Drishadvati valleys up to the Bahawalpur (Pakistan) border and the first site in Uttar Pradesh was reported at Alamgirpur in Meerat in 1958-59. A Harappan site near Kotla Nihang Khan, Ropar, was excavated in 1950s. The basic stratigraphic and ceramic index of the situation in the Ghaggar

Valley section of Rajasthan was provided by the Kalibangan excavations which continued through 1960s, and by the early 1970s, the Kalibangan index was related to the sites of Haryana. Banawali, Rakhigarhi, Balu and Bhirrana in Haryana and Hulas and Alamgirpur in western Uttar Pradesh are among the major Harappan sites excavated in this region since then.

Kalibangan

Of the three mounds of Kalibangan (Lal 1979, Bala 2004) on the left bank of the dried-up Ghaggar course-the higher western citadel mound, the lower eastern and exclusively residential mound, and a much smaller area with only fire-altars to the east of the eastern mounda detailed evidence of planning has emerged. It is not just regarding these mounds but also at the cemetery site to the west-southwest of the western mound. The thick mud brick fortification wall ( 3 m to 7 m thick) of the citadel mound had two structural phases and salients or bastions placed at regular intervals. The mound itself ( $c$. $240 \mathrm{~m} \times 120 \mathrm{~m}$ ) was divided into roughly equal northern and southern sectors by an equally thick wall. The salient placed in the centre of the southern side was 17 m wide and projected 9.35 m from the main wall face. It had an extension as well, and the extension was battered on the outside. The southern sector of the citadel mound was entered from outside the fortification by steps leading to a passage. The ascent of the steps was protected on both the sides by a 2.6 m wide structure with salients. To the north, there was another entrance to the southern sector of the citadel mound through a mud brick staircase located between two centrally located bastions on the dividing wall. There were two entrances to the northern sector as well, one at the northwestern corner and the other at the end

Fig. 27. Banawali site plan.
of the eastern section of its wall. The northwestern entrance was marked by a Vshaped area between the two bastions. There were houseblocks in the northern sector and they were separated from the dividing wall by a wide passage paved with bricks on edge. There are also signs of a road in the northern sector.

The southern sector showed only five or six high, differently sized mud and mud bricked platforms, all with passages of different widths between them. Access to the top of these platforms was provided by stairs. Whatever occurred on the top of these platforms was found mostly destroyed but in one case the evidence was clear.

A atop one of the platforms there lay a series of seven ‘fire-altars’ in a row. Behind these firealtars ran a wall in a north-south direction, which shows that people had to face the east while performing rituals at these altars. The altars were
oblong on plan, sunk into the ground and lined with clay. They contained ash and charcoal, besides a cylindrical and faceted clay (burnt or unburnt) stele standing up near the centre. Though in the series under discussion only fragments of what are called ‘terracotta cakes’ were obtained, elsewhere these were found in sufficient numbers showing that they formed some kind of an ‘offering’. To the west of these fire-altars lay embedded the lower half of a jar. It contained ash and charcoal and was evidently connected with the use of fire-altars. (Lal 1979; Tarkhanewala Dera, Trivedi and Patnaik 20032004).

The southern sector has a well, the only well found at Kalibangan, and a few pavements. Apart from the platforms mentioned above and these finds, there is no other structural find in the southern sector of the western mound of Kalibangan. A burnt-brick drain, again the only example of its kind in Kalibangan, apparently served as the means of sewage in this sector. The excavated pavements here might represent bathing places. People would bathe in the water drawn from the well and then ascend the stairs to the top of the platforms to perform their rituals. It is only a possibility but a rational inference on the whole. The archaeological evidence suggests that these rituals involved the lighting of fire in sacrificial pits.

Between the western and the eastern mounds, there is an open space at the site. The eastern mound, which measured $c .240 \mathrm{~m} \times 360$ m , was also fortified by a 3 to 3.9 m wide wall with bastions at the four corners. This was constructed in a box-pattern with mud-filling inside mud brick veneers. In the central part of the western side, there was possibly a 3 to 7 m wide gateway with a guardroom attached to it. The eastern side also might have a similar entrance. There is a third entrance in the

Fig. 28. Rakhigarhi site plan.
northwestern corner, and in the southeastern corner there was possibly a fourth entrance. The planning of the eastern mound included at least 5 major north-south roads and an assortment of lanes, some of which were occasionally paved with terracotta nodules and had soakage jars. A wooden drain has been traced by the roadside at one spot. There are also small mud brick platforms by the side of the roads. Their purpose is uncertain. A corridor from a lane usually led to a courtyard with $3 \mathrm{~m} \times 2 \mathrm{~m}$ or $2 \mathrm{~m} \times 1 \mathrm{~m}$ rooms disposed around it. The floors were made dampproof with a soling of terracotta nodules and bits of charcoal. In one case, the floor of a room was found paved with tiles decorated with intersecting circles. Only one socket has been noticed on the door sills and this suggests the prevalence of $c .70$ to 75 cm wide single-leaf doors. There were fire-altars or sacrificial pits in many of the houses and these are identical with the examples found on the top of the platforms of the southern sector of the western mound. Troughs made of mud brick, which have been found inside the house limits, may suggest the way of offering fodder to the domestic
animals. A staircase found in a house implies that double-storeyed houses were not unknown. The roofs were likely to be flat and made of mud. In a structure to the east of the eastern mound, five fire-altars, and nothing else, have been excavated.

Tarkhanewala Dera is a site near Kalibangan and excavations here have led to the find in the residential area of a fire-altar lined with mud bricks and full of charcoal, ash, terracotta cakes and potsherds. After the structures fell into disuse, the area was occupied by potters’ kilns, two of which have been excavated.

BanaWali

On the right bank of the Ghaggar between Hissar and Sirsa, Banawali (Bisht 1982,1987, 1998-99) shows a low mound which was extensively ploughed over by the villagers. The mound, an irregularly shaped rectangle of $c$. $300 \mathrm{~m} \times c .250 \mathrm{~m}$, measures about 9 ha in extent.

On the southern side, about a third of the area stands for ‘acropolis’ separated from the lower town by a roughly arch-shaped and c. 5.4 to 5.47 m thick fortification wall. The evidence of a 5.7 m to 6.5 m wide and 3.6 m deep moat along the eastern side of the fortification wall is clear. It may be surmised that the moat encircled the site on the other sides as well. Harappan structures have been traced primarily in the eastern section of the fortified area.

An important part of the Banawali planning is that there was a “peripheral road running all through uninterruptedly on either side of the fortifications” of both ‘acropolis’ and ‘lower town’. The main entrance-an elaborate gate complex-lay on the east, comprising “a frontal moat, flanking bastions, a broad passageway, a postern stairway, and a storm-water drain”. The plans have not been published in detail and thus

the foregoing details do not make any clear sense.

The streets of the town vary in width between 4.20 m and 9.10 m . Seven such streets and two lanes have been traced. One large ( 52 m $\times 46 \mathrm{~m}$ ) house was approached from a lane and this was found to possess a large courtyard, a bath with traces of a soakage jar and a burnt-brick-built drain. Plenty of ash and charcoal was found in one of its rooms which showed a platform built against one of the walls and some ‘fire-places’ in front of this platform. A terracotta block was found standing in the centre of one of these ‘fire-places’. On the Kalibangan analogy, these ‘fire-places’ involved rituals with the use of fire.
“A rich harvest of seals, weights, beads, including those of gold, lapis and etched carnelian” and good quality pottery came from another house-complex which also had a room with a ‘fire-place’. In another room of this house, which lay next to a road, there were many jars, all half-embedded in the floor.

The analysis of wood charcoal obtained in the excavations showed trees which still extensively grow in the area. “The tree and shrub taxa represented by wood charcoals from Banawali are the constituents of deciduous forest, characteristic of arid and semi-arid regions”.

Balu and Rakhigarhi

Balu is in the Kaithal district of Haryana and excavations here have revealed a fortified Mature Harappan settlement ( 108 m x 96 m ). The width of the fortification wall is about 12 m . There is an entrance in the southeastern corner of the fortification. The houses were planned alongside streets and lanes ( 1.45 m to 1.90 m wide). A drain of burnt bricks has been found
somewhere along the streets. The evidence of fire-places was clear, one particular example being elliptical ( $1.69 \mathrm{~m} \times 96 \mathrm{~cm}$ ) in shape with burnt or unburnt clay ‘cakes’ set to form the fireplace. Carbonised wheat/barley grains were found scattered around it and inside there were charred bones.

Supposedly on the right bank of the dry course of the Drishadvati in the Hissar district of Haryana, Rakhigarhi shows a cluster of seven mounds marked as RGR-1 to RGR-7 covering a total area of about 105 ha , thus making it a site comparable to all other so-called ‘metropolitan’ sites of the Harappan Civilisation. Amarendra Nath, who began work at the site in 1997-98, thinks that these mounds, except RGR-4 and RGR-5 which are connected, may stand for separate zones of individually fortified occupations. RG-5, about 17 m high, seems to be the highest mound. RG-2 has been considered the citadel mound but the logic of this idea is not clear to me. Its fortification wall, although made of mud bricks, possessed a veneer of burnt bricks both on the inside and outside. Mud brick platforms of the type found in the southern sector of the western mound of Kalibangan have also been noticed in the Rakhigarhi citadel (i.e., RG-2). There were streets on all sides of one of these Rakhigarhi platforms and a covered verandah with a flight of steps provided approach to its top. Rectangular pits show signs of burning and presence of bones inside. A brick was vertically placed in the centre of the pit in one case. The shape of these pits was usually rectangular but there were other shapes as well: heart-shaped in one case, semi-circular central projection on the north and south sides in case of another… The published details of the excavations are not yet particularly clear. A large bead-making workshop has been located along one of the

Fig. 29a. Layout plan of Lothal.

Fig. 29b. Reconstruction of the Harappan city of Lothal.

Fig. 30. Lower city, Lothal.

Fig. 31. Warehouse platforms, Lothal.
streets of RGR-1 with 3000-odd finished and unfinished beads of carnelian, agate and jasper and also many bead-polishers and hearths for heating the stones at different stages of manufacture. Platforms and north to south streets have been noticed in RG-4, whereas RG-5 is supposed to show bone and ivory crafting. RG7 has been interpreted as a burial site.

BHIRRANA

Period II was the transitional phase the Mature Harappan development at Bhirrana in the Fatehabad sector of Haryana. The settlement was

Fig. 32a. Bathing platform and attached drain, Lothal.

Fig. 32b. Row of bathing platforms and attached drains, Lothal.
fortified during this period and it was during this very period that planned streets and mud brick houses made their appearance with dish-on-

stands, button-based goblets and a hoard of 3061 semi-precious stone beads comprising many specimens of lapis lazuli, carnelian, faience, steatite and shell.

In the succeeding Mature Harappan period, the mud brick fortification wall was 2.15 m to 3.75 m wide and had a gateway at the southwestern corner. The walls of the houses are said to be 24 degrees away from the true North. Such houses possessed 3 to 4 rooms, drains, hearths/ovens, brick-laid bathroom floors and storage pits. Circular hearths filled with ash and broken terracotta crucibles were found outside the western fortification wall, and as one of the crucibles was found to have had a bit of copper sticking to it, the whole complex has
been interpreted as a coppersmithy. One street ran along the fortification wall and another street has been traced in the centre of the settlement. The pottery of this phase has been called classical Mature Harappan. Copper occurs widely, including Ganeshwar-type arrowheads and two inscribed copper celts. Spokes have been found painted on some terracotta wheels. Gold beads occur along with five typical Harappan seals, an ithyphallic terracotta and a female terracotta figurine.

MISCELLANEOUS SITES OF THE REGION

In Haryana, Indian Punjab and western Uttar Pradesh, Harappan sites are wide-spread. Although some of them have been excavated

Fig. 33. Dholavira house-complexes.

Fig. 33a. Layout of Dholavira.
(cf. Rohira), virtually in all cases nothing more than a stratigraphic succession is obtained, and it is difficult to ascertain if the excavated Harappan levels belong to the Mature or Late phases of the civilisation. The results of excavations at Alamgirpur and Hulas, both in western Uttar Pradesh, offer a good illustration of the problem. At Alamgirpur, the beginning of occupation at the site seems to go back to $c$. 2500 вс because the date of about 2200 вс has been obtained from a level about 1 m higher than the earliest level. Chronologically,

Alamgirpur which has also yielded evidence of an inscribed seal should have a level identical with the Mature Harappan.

Some doubts exist about the Mature Harappan status of Hulas (Dikshit, 1984)as well. Its ‘middle’ and ‘upper’ phases belong to the Late Harappan, but the pottery with classic Mature Harappan motifs, such as pipal leaves and peacocks, suggests that its ‘early’ phase is likely to be Mature Harappan. The structural evidence of this phase comprises the ruins of rectangular mud brick structures including two

Fig. 34. One of the main entrance gates at Dholavira.
complete rooms ( $3.10 \mathrm{~m} \times 2.40 \mathrm{~m} ; 3.10 \mathrm{~m} \mathrm{x}$ 3.40 m ), a storage pit and hearths. Interestingly, one of the radiocarbon dates from the site is around 2500 BC uncalibrated, which would mean that its calibrated or true historical date is about 3000 BC . The point is that there is no reason to doubt offhand the Mature Harappan status of many sites of Haryana, Punjab and western U.P. More excavations and more dates are needed.

GUJARAT

Gujarat possesses Harappan sites of all phases from the Rupen estuary in the north to the Tapti estuary in the south and also in the peninsulas of Saurashtra and Kutch, including the bets or islands of the Rann of Kutch. M.S. Vats, who wrote the main Harappa excavation report, identified Rangpur as a Harappan site in the 1930s, and by the 1950s and 1960s the importance of the Harappan Civilisation in the archaeology of Gujarat was well understood. Excavations have continued since then at many sites including Surkotada, Dholavira, Juni Karan, Nagwada, Padri, Kuntasi, Bagasra, Kanmer, Kshirsara, etc.

LOTHAL

A 13 m to 22 m wide and 1.8 m to 2.4 m high wall enclosed the settlement of Lothal (Rao 1979) which was supposed to have been reached by an arm of the sea from the Bhogavo creek. A burial ground lay to the west of this walled settlement ( 240 mx 210 m ) whose entrance was in the south. The houses of the south-eastern segment of the settlement stood on higher mud and mud brick platforms in relation to the houses of the other areas, and this is the reason why

Fig. 35a. Use of dressed stone in architecture, Dholavira.

Fig. 35b. Dressed pillars and bases found in the excavations, Dholavira.

the three excavated blocks of houses of this section have been supposed to belong to an ‘acropolis’ by the excavator S.R. Rao. This ‘acropolis’ was in contrast to the ‘lower town’ outside it. One particular houseblock has been called a ‘warehouse’. It stands on a mud brick podium (c. $49 \mathrm{~m} \times c .41 \mathrm{~m} \times c .4 \mathrm{~m}$ ) with an outer platform. This podium had on it 12 solid blocks of mud brick (each 3.65 m square and 0.91 m high) separated by $c .1 \mathrm{~m}$ wide crisscross passages. These blocks presumably carried wooden superstructures, and 65 terracotta sealings with impressions of reed, woven fibre, matting and twisted cords were found in this complex. Another houseblock in ‘acropolis’ shows a row of 12 houses (each $7.62 \mathrm{~m}-8.53$ m long and 5.48 m wide) which had their individual bathing pavements connected with a drain outside. A house in the ‘lower town’ was found to have three rooms fronted by a 12.19 m wide verandah. Bathrooms connected with the outside drains are a common feature in this part of Lothal. Small sacrificial pits containing terracotta cakes or round clay lumps and ashes are common. The streets ( 3.65 m to 5.48 m wide) and lanes ( 1.82 m to 2.74 m wide) were paved

Fig. 36. Singboard which may have been on the North Gate.
with mud bricks, with a layer of gravel on top. The street drains (internal width 1.21 m to 0.64 m ) were arranged with a gradient which led to cess-pools, one of which was 1.31 m square and c. 1.52 m deep. The famous dockyard of the site ( 218.23 m on the west, 215.03 m on the east, 35.66 m on the south and 37.49 m on the north) was enclosed by a four-course burnt-brick wall set inside a broader mud brick embankments. There were two inlets and a spillchannel, which could regulate the level of water inside. Its identification as a dock has been proved by the presence of marine organism in it. An open mud brick platform ( 12.8 m x 243.84 m ) adjoining the western embankment of this ‘dock’ has been interpreted as a loading platform. There is also a hypothesis that the level of the sea was higher during the period of the Harappan Civilisation than its present level and thus Lothal was directly accessible from the sea.

SURKOTADA

This settlement (Joshi 1990) is on the model of the two sectors of the western mound of Kalibangan. A wall divided it into two units, possibly one residential and the other nonresidential. An artificial mud platform in the nonresidential section was 1.5 m high whereas in the residential section the height of such a platform was 0.5 m . The non-residential area was 60 m square and surrounded by a 7 m wide mud and mud brick fortification wall (rubblecovered on the outside and mud-plastered inside) with a mud brick buttress on the east. The wall around the residential section was less than half the thickness of the wall around the non-residential area. It was 3.25 m wide. The non- residential section was entered from the south whereas the residential section was entered from the southeast. The burial ground

lay to the northwest of the fortified area. The excavated buildings do not deserve notice.

Dholavira

Stage IV marks the Mature Harappan occupation at Dholavira (Bisht 1989) which is located in an island of the Rann of Kutch and measures about 100 ha in extent, with its outer fortification enclosing an area of about c. 781 m x c. 630.50 m . According to the excavator R.S. Bisht, “almost all the salient features of the city-planning were scrupulously maintained along with the monumental structures such as gateways, fortification, drainage system” during this occupational stage of the site. There were regularly spaced bastions along the outer wall: 11 in the northern wall and 9 in the western wall, all at a mutual distance of 50 to 52 m . The bastions have not been fully traced in the other walls. A small opening in the southern section of the outer wall is the only such opening which has been traced but there must be similar openings elsewhere in the wall too. The outer wall, made of mud brick with a stone veneer on the outside, was 8.40 m wide in the southeast. It has not survived in any other section. The only area which was directly protected by the outer wall is the ‘lower town’. There are three properly fortified areas inside the outer wall: a citadel area consisting of a high ‘castle’ with the eastern adjunct of a lower ‘bailey’, both fortified, and a ‘middle town’, also within a perimeter wall strengthened by bastions. The external measurement of the castle was 118 m x 151 m , whereas the external measurement of the citadel as a whole (castle and bailey) was c. 140 $\mathrm{m} \times \mathrm{c} .280 \mathrm{~m}$. The internal measurement of the middle town was $242 \mathrm{~m} \times 340.5 \mathrm{~m}$. The occupational deposit in the castle was 14 m thick, whereas the occupational deposit in the bailey, an area said to have been never occupied for

Fig. 37. One of the water tanks excavated at Dholavira.
regular living, was 7 m . The castle had a fortification whose extant height is 9 m and width 15.5 m . The fortification was steeply battered and had gates on all sides, two of which have been excavated on the north and the east. The eastern wall had two gates, thus making the total number of ‘castle’ gates five. The major excavated eastern gateway was fronted by a terrace, from which the entry passage led to a wide corridor flanked by a chamber on the south. The corridor led to a flight of stairs up the wall. The level of the flanking chamber is said to be higher than the level of the corridor and its open front was supported, as seen on the plan, by three pilasters with polished limestone blocks as base. The stairs led to an upper terrace between two bastions, and from there another series of steps led to the ground level. Between the entry passage and the inner corridor there

Fig. 37a. Artist’s impression of Dholavira city.
are signs of a door-sill made of large limestone slabs.

The north gate of the castle has been said to be the ‘most majestic’ of the Dholavira gates. Here the inner corridor is flanked by two chambers, both elevated above the level of the corridor and possessing open fronts. The corridor led to a terrace which was connected with a ‘ceremonial pathway’. The limestone sockets of the door- sill survive and suggest a heavy doorframe with two door-leaves between the corridor and the terrace. The corridor has a similar door at its incoming end. The chambers on the two sides are likely to have had three pillars supported by stone blocks, as was the case in the eastern gate, but they have not survived.

The inner structural details of the castle and the bailey do not seem to be available. All that we seem to know about the bailey is that it had a number of gates, especially one on the north where a flight of steps led to a corridor ( 7.30 m x 2.30- 2.55 m ) flanked by ‘elevated’ chambers.

A section to the south of the ‘castle’ has been marked as ‘annexe’ on the published plan. This was ‘meant for housing the retailers and menials’ who served the people living inside the castle complex. More interestingly, the castle-bailey complex was fronted on the north by an open space ( $283 \mathrm{~m} \times 45-47.50 \mathrm{~m}$ ) which ended at the wall of the ‘middle town’. Along the wall of the castle-bailey, there were 4 long terraces arranged in an ascending order. This was apparently a sitting arrangement and gives the

impression of this open space being used as a stadium. This large stadium had to its east another small stadium. The structural details of the fortified ‘middle town’ and open ‘lower town’ are not available, but the alignments of the roads seem to have been shown on the plan.

Two places on the plan, one to the south of the ‘castle’ and the other to its east, have been marked as water reservoirs. The larger one lay between the ‘castle’ and the ‘annexe’ and was at least 95 m long and connected by a spillchannel with another reservoir which is still unexcavated. The width of this reservoir at the top was 11.42 m and 10.80 m at the bottom. The depth varied from 1.95 m to 4 m because the sloping bedrock at this place was left undisturbed. The reservoir to the east of the ‘castle’ was at least 70 m wide (north-south) and a flight of 31 steps led to the bottom. Both these reservoirs were partially cut into solid bedrock. It has been said that the Harappans created at least 16 water reservoirs within the outer walls of Dholavira. Of different sizes, they were arranged “along the northern and western and largely along the southern sides of the main settlement and to the east of the citadel”. The reservoirs occupied a total area of about 10 ha . Two more features of water management in the city area deserve notice. First, there was an intricate drain network inside the citadel area to lead rainwater “to a reservoir carved out in open space provided in the bailey”. Secondly, in the south-western corner of the castle, there were two water tanks and a well, the latter a little more than 4 m in diameter.

On the basis of the debris scattered around, it has been inferred that water was possibly drawn from this well in leather bags tied to ropes which were manipulated by a system based on a wooden pulley. The water thus collected was stored in a large tank with a floor of polished
stone. Its sides were partly lined by limestone blocks and it had a superstructure of dressed stone. All things considered, the water management system of Dholavira is truly amazing, although it is not yet fully studied from the technical point of view. The two riversthe Mansar in the north and Manhar in the south, the latter cutting through the southeastern section of the site-were controlled by check dams at a number of points and the water was diverted to the reservoirs inside the walls, among other places.

The cemetery area lies to the west of the outer wall, but not much is known about it.

Dholavira is possibly the most complex of all Harappan Civilisation sites: the maze of ‘castle’, ‘bailey’, ‘stadium’, ‘middle town’, ‘lower town’, the spread of about 10 ha of reservoirs, and the elaborate system of gateways. Its detailed plan does not seem to have been published yet. Dholavira seems to lie on a straight route across the Great Rann to Sind.

The concern with water management that one observes at Dholavira clearly carries the message that the climate of Kutch was as arid then as it is now-about 10-15 inches of annual rainfall with frequent monsoon failures. The extensive grazing lands, raw materials like chert and various semi-precious stones, and rich potential for the cultivation of cotton might have been among the factors to attract the Harappans to Kutch. The Banni area which juts like a peninsula into the northern Rann is a rich pasture land and Kutch in general is noted for its breed of horses and cattle. It is extremely unlikely that Dholavira was a maritime port. Apart from the geographical factor, there is nothing in the material found at the site which would suggest external links, maritime or otherwise. The only thing one can recall is a fragment of a chlorite

vessel with mat motif. Mohenjodaro also yielded a specimen like this, and although it is easily related to similar finds over a large part of west Asia, its occurrence at Dholavira proves in no way that the site was a maritime port.

Kuntasi and Other Fortified Sites in Gujarat

Kuntasi (Dhavalikar, Raval and Chitalwala 1996; Rao 1963; for Padri, Shinde 2004; for Juni Kuran, Pramanik 2003-2004) is located on the bank of a small river at the eastern edge of the Gulf of Kutch. It had two chains of fortifications made of boulder and mud, with space of 15 m to 25 m between them. The site also possessed a watchtower overlooking the river, a gateway, a multi-roomed structure with kilns, furnaces, storage areas, and finally, a ramp going down to the river. The site seems to be a workshop of pottery, beads and copper and was also a river port linked to the sea. Desalpur in Kutch is a fortified settlement but very little of the site was excavated. Meanwhile, Kutch was an important zone of fortified settlements. R.S. Bisht refers to the discovery of 60 Harappan sites in this region, of which 70 per cent are supposed to belong to the Mature phase. About 11 of these 60 sites “have shown evidence of fortification while many others provide strong suggestive indication of the same”. In this context, a mention may be made of Khirasara in the Nakhatrana area of Kutch, where Desalpur also is located. In Saurashtra, Bagasra on the southeastern coast of the Gulf of Kutch is a fortified shell workshop site and the same is roughly true of Kanmer and Shikarpur, fortified with craft-working areas. At Shikarpur, the fortified area was spread over about 1 ha and its fortification wall was about 10 m thick. There was an open space in the centre with small structures surrounding it. Shikarpur lies on a creek of the Gulf of Kutch. Bagasra or Gola

Dhoro in its Mature Harappan phase had a 7 m thick fortification wall, with the fortified area being about 100 m square. Settlement spread outside the fortified area also and yielded in that sector steatite seals and miscellaneous evidence of craft activity.

JUNI KURAN

The site lies roughly 42 km to the northwest of Dholavira and is roughly $410 \mathrm{~m} \times 350 \mathrm{~m}$. In the Mature Harappan context, the site had a fortified citadel complex within an outer fortification. This citadel is located in the northern part of the enclosed settlement area. The citadel area measured $72 \mathrm{~m} \times 92 \mathrm{~m}$ whereas the outer fortification enclosed an area of $220 \mathrm{~m} \times 225 \mathrm{~m}$. The walls were made of rubble set in mortar. The gateway in the southwest section was approached by a 3.40 m wide corridor and originally the entrance was 6.42 m wide. Another gateway was traced further east along the fortification wall, with a set of large chambers between them. The second gateway led to a spot described as Stadium I which had, apparently at its eastern edge, Stadium II. Stadium I shows a brick platform (c. 25 m long and 8.25 m wide) which slopes from the side of the chambers set inside the fortification to the ground in front. This is supposed to be common people’s stadium whereas Stadium II ( 22.20 x 14.50 m ) is marked by another sloping platform, and this stadium is supposed to be a stadium used by the elite. In between the citadel and the structures of the lower town is an open area of $200 \mathrm{~m} \times 50 \mathrm{~m}$.

SOME GENERAL OBSERVATIONS ON THE
FEATURES AND DISTRIBUTION OF HARAPPAN SITES

Looked at closely, the major Mature Harappan settlements display considerable variations in

Fig. 38. Wells at Mohenjodaro.
planning. No two settlements are exactly alike, despite the presence of several common features. If one takes the case of the three or four largest settlements excavated so farMohenjodaro, Harappa, Dholavira and Rakhigarhi-the variations in the details of their planning would be clear. According to the excavator of Rakhigarhi, several mounds of the site were all individually fortified, with no common outer wall. Dholavira had separate fortified components, but all of them-the castle, the bailey, the middle town, the lower townlay within one common outer wall. The distinction between the castle and the bailey of Dholavira, does not seem to be repeated elsewhere. It has been noted that about 10 ha area of this 100 ha site was given over to the
water reservoirs, the base of some of which was apparently dug into the bedrock. In fact, the water management system of Dholavira where the water held by check dams outside the fortifications could be guided into the reservoirs within the fortified area, seems to be a feature unparalleled elsewhere. Even the oft-quoted division between the ‘citadel’ mound and ‘the lower town’ of a Harappan city is not as straightforward as it sounds. First, this division is not universal. Among other things, the distinction between the acropolis and the lower town is blurred in the case of Banawali and Lothal and all the other recently excavated Harappan fortified settlements in Gujarat. Secondly, if one looks at the planning of Surkotada, it is apparent that its planning is based

Fig. 39. Wedge-shaped bricks used in wells.
on the model of the western or citadel mound of Kalibangan. No such citadel mound exists at Surkotada. At Lothal, only one particular sector has been given the status of an ‘acropolis’ simply because the platform bases of the houses in that sector were higher. It was not a citadel in the sense of the Mohenjodaro, Harappa and Kalibangan citadels. Again, the features like drains, wells, and even streets and lanes vary in considerable detail from site to site. No site has shown as many drains and wells as Mohenjodaro, and the only site which has suggested a similar easy distribution of drains and wells is Lothal. Being a much smaller site than Mohenjodaro, the occurrence of drains and wells at Lothal is proportionately less but in view of its size, abundant.

Despite all these, and perhaps many more variations in detail, what cannot be ignored is the presence of a normative concept of planning and the technical ability to execute it. At Mohenjodaro, for instance, apart from earmarking the sites of the two mounds, they had to focus on the construction of two massive and high platforms as the base of the numerous smaller platforms of uneven heights, on which the individual buildings would be built. The drainage system was planned and meant to
cover each and every part of the city. The location and planning of the important buildings on the top of the western mound had to be visualised right at the beginning and the fortification systems of both the western and the eastern mounds would also have been among the initial concerns of the town-planners. The alignments of the roads, streets and lanes and the laying of the houseblocks must all have figured in the conceptual framework of the Mohenjodaro planners. None of these things would have been possible unless the planners had a clear concept of the type of city they were going to build. One would argue that in the case of Mohenjodaro all these things seem to be the result of a conscious decision in the light of the planning norms which apparently evolved in the Early Harappan period at other sites in other areas. At Mohenjodaro, this initial phase is absent, and we assume that Mohenjodaro was built as a single planned unit after the lessons of city-planning were mastered elsewhere.

The Mohenjodaro image in this sense contrasts sharply with what we learn at Dholavira. At Dholavira, the planning of the Mature Harappan city was given an almost complete shape by the end of the Early Harappan stage of the site. The excavator, R.S. Bisht, has always emphasised this point. A moat has been identified at Banawali and Jansen has argued for the existence of a moat at Mohenjodaro. At these two places at least we get the image of a walled city surrounded by a moat, an image which the planners of the early historic Indian cities would have well understood. Perhaps more important is the presence of water reservoirs inside the fortified limits of Dholavira. It is possible to argue that cities like Mohenjodaro and Dholavira embodied the concepts of town planning that one finds enshrined in the later historic cities of the

Fig. 40. Street at Mohenjodaro with covered drain.
subcontinent. It is regrettable that very little is published in detail about Dholavira, but the fact remains that the lay-out of this city situated in an apparently water-starved area of the country shows resilience in planning, especially regarding its water management. The concept of city-planning in the Harappan Civilisation was not something monotonous and regimented but was flexible enough to adapt to regional contexts and evolve a new ideal. The town planning norms were also backed by a degree of technical competence which we observe at its best in the water management systems of Mohenjodaro and Dholavira.
M. Jansen (1989) has made a special study of the water-management system of Mohenjodaro with reference to its wells and drains. The cylindrical wells of Mohenjodaro were made of specially prepared wedge-
shaped bricks. Made of specially prepared wedge-shaped bricks, the wells of Mohenjodaro were circular in form, and it has been claimed by Jansen, who has made a special study of the Mohenjodaro wells and drains, that the circular form of the wells was "statically ideal for withstanding the lateral pressure exerted on the shafts which were sunk to a depth of at least 20 m ". It is important to note that no well of this kind has been found in the Early Harappan level anywhere and thus there is reason to claim that this form of wells was invented during the Mature Harappan period. Water supply through cylindrical brick-built wells was unknown in Egypt and Mesopotamia. The ‘only technically feasible manner of construction’ of the Harappan wells would be the shaft-sinking method which is still used in the region. On the basis of number of the excavated wells, it is presumed that the

Fig. 41. Drain with round curve at bend, Mohenjodaro.
"radius of the mean catchment area supplied by each well in Mohenjodaro, ‘was a mere 16 m .’ The wells usually lay within a house but occasionally they were placed between two houses, and in two such cases “brick-benches were built around the well for the use of people awaiting their turn to draw water”. The abundance of wells at Mohenjodaro is striking, and the only possible reason is that the water of the Indus was then far enough for the inhabitants to use. The place of wells has been taken up by water reservoirs at Dholavira and at Kalibangan
the inhabitants could apparently depend on the water of the Ghaggar. The western mound of Kalibangan has a single well, the lone well in the entire settlement. It is possible that the water of this well was meant to be used only for ritual cleaning and bathing.

Jansen has described the network of burnt-brick-built drains with differing depths and widths as “the most impressive engineering feat accomplished by the Harappan Civilisation”. The slope of the drains was about 2 cm per metre.

Wherever a drain had to traverse a long distance or where several drains met, a brick cesspit was installed, this being the simplest method to avoid clogging by allowing solids to settle. The effluent flowed into such a brick shaft at a high level, filled it and flowed out from the other side at a slightly lower level. The suspended matter gradually formed a deposit which could be removed by steps leading down into the pit, which was presumably covered by a loose wooden roof.

Open soakpits were also commonly used, especially at the junctions of small lanes and bigger streets, and had to be regularly cleaned. When the street drains were far from the houses, vessels were placed under the vertical chutes placed on the outside walls of such houses. These soakage vessels also needed regular cleaning. Some vessels had perforated bases and served the function of cesspits. The housedrains were usually connected to the public or municipal drains. What, however, is interesting is that the vertical chutes coming from the houses ‘very often’ stopped short a little above the public drain or the receiving soakage jar. This implies that the effluent coming out of the houses was open to the air for a while, which must have naturally caused bad smell. This also meant that regular cleaning of the soakage jars

and a constant high flow of water through the drains were very necessary. Jansen writes, “Both measures imply a large workforce engaged in fetching water for flushing and in clearing out deposits”. This was indeed reminiscent of the municipal workforce for cleaning that one could observe in many Indian municipalities till fairly recently. Bathing platforms were also unique to the Harappan Civilisation. More examples have been found at Mohenjodaro than at any other Harappan site. They occurred either in small rooms meant for the purpose or in larger rooms where an area was kept separate for it. Their edges were raised, making them a bit like a basin and they sloped towards the drains out of the rooms. A toilet was often placed in the outside wall of the bathing platform, sometimes with its own vertical chute. Two raised side-brackets held the seat. Other toilets consisted simply of a hole in the floor immediately above the chute. With modifications, this is the general picture of a domestic Indian bathroom without the provision of running water.

A very important implication of this study of Mohenjodaro drainage by Jansen is that there was a municipal or public system of cleaning the drains, cesspits, soakage vessels and the like.

Whether water was drawn from the wells of Mohenjodaro in leather bags through a pulley system is not known, but at Dholavira this has been suggested to be the case. Considerable engineering skill must have been involved in laying out a system of drains to lead rainwater to a storage system in the ‘bailey’ part of the Dholavira citadel. The fact that an area of about 10 ha was covered by water reservoirs within the city walls of this site and the fact that they were partially cut deep into the bedrock must be considered a marvel of technical expertise. In an area of high summer temperature, low
rainfall and frequent monsoon failures, the creation of such waterbodies was something far more than mere necessity. In the blistering hot Kutch summers, it must have made the city look restful and composed in contrast to what lay outside.

The drains, bathrooms and toilets did not occur on the same scale everywhere. At Kalibangan, it was not generally prevalent and perhaps toilets were totally absent at the site. Streets and lanes have been noticed at about all excavated sites but street pavings of any kind were rare. Their width naturally varied-from about 11 m in one instance at Mohenjodaro to about an average of 1.8 to 2 m in the case of lanes at many other places. Width also varied from one part to another in the same street. Such variations notwithstanding, the streets were by and large straight, with eyes moving down from one end to the other in one sweep. The lanes could twist and turn but the corners were usually right-angled. The lay-out, even of the lanes, was obviously worked out before.

As far as the distribution of sites is concerned, one has to take note of the fact that the sites were well integrated into a vast and varied landscape. Baluchistan is very different environmentally than the foothill areas of Punjab, Haryana and western U.P. However, the Harappan sites everywhere seem to be deeply rooted in the land. The factors behind the locations of Harappan sites still remain to be worked out in detail, but when the details of many such locations are available, that would possibly lead to a better understanding of the individual features and planning.

Regarding the analysis of different units of planning at Mohenjodaro, an important study is that by Mohan Pant and Shuji Funo (2005) who have shown that the house blocks of

Mohenjodaro, Bhir mound, Taxila and the town of Thimi in the Kathmandu Valley were all

REFERENCES

Bala, Madhu. 2004. Kalibangan: Its Periods and Antiquities. In D.K.Chakrabarti (ed.), Indus Civilisation Sites in India-New Discoveries. Mumbai: Marg Publications, pp. 34-41.
Besenval, R. 1994. The 1992-1993 Field-seasons at Miri Qalat: New Contributions to the Chronology of Protohistoric Settlement in Pakistani Makran. In A.Parpola and P. Koskikallio (eds.), South Asian Archaeology 1993. Helsinki: Suomalainen Tiedeakatemia, pp. 81-92.
-. 1999. The Protohistoric Site of Shahi-Tump in Pakistani Makran. South Asian Archaeology 1999 (unpublished).
Biagi, P and M. Cremaschi. 1990. Geoarchaeological Investigations on the Rorhi Hills. In M.Taddei and P. Callieri (eds.), South Asian Archaeology 1987. Rome, ISMEO, pp. 31-42.
Bisht, R.S. 1982. Excavations at Banawali: 1974-77. In G. Possehl (ed.), Harappan Civilisation, Delhi: Oxford and IBH, pp. 113-24.
-. 1987. Further Excavations at Banawali: 1983-84. In B.D. Chattopadhyay and B.M. Pande (eds.), Archaeology and History. Vol. 1. Delhi: Agam Kala, pp. 135-56.
-. 1989. The Harappan Colonisation of the Kutch: An Ergonomic Study with Reference to Dholavira and Surkotada. In K.Deva and L.Gopal (eds.), History and Art. Delhi: Ramanan Vidya Bhavan, pp. 267-72.
-. 1998-99. Dholavira and Banawali: Two Different Paradigms of the Harappan Urbis Forma. Puratattva 29: 14-37.
Dales, G.F. 1962. Harappan Outposts on the Makran Coast. Antiquity 6: 86-92.
-. 1965. New Investigations at Mohenjo-Daro. Archaeology 18: 145-50.
-. 1979. The Balakot Project: Summary of Four Years of Excavation in Pakistan. Man and Environment 3: 45-53.
planned by using Kautilya’s Rajju and Danda units of linear measurement.

Dales, G.F. and J.M. Kenoyer . 1986. Excavations at Mohenjo-Daro, Pakistan: The Pottery. Philadelphia: University Museum.
-. 1992. Recent Excavations at Harappa. Eastern Anthropologist 45: 21-37.
Dhavalikar, M.K., Raval, M. and Chitalwala, Y. 1996. Kuntasi, a Harappan Emporium on West Coast. Pune: Deccan College.
Dikshit, K.N. 1984. The Harappan Levels at Hulas. Man and Environment 8: 99-102.
Fairservis, W.A. 1982. Ahladino, an Excavation of a Small Harappan Site. In G.L. Possehl (ed.), Harappan Civilisation-A Contemporary Perspective. Delhi: Oxford and IBH, pp. 107-12.
Flam, L. 1986. Recent Explorations in Sind. In J. Jacobson (ed.), Studies in the Archaeology of India and Pakistan. Delhi: Oxford and IBH, pp. 65-89.
Franfort, H.P. 1984. The Harappan Settlement of Shortughai. In B.B.Lal and S.P.Gupta (eds.), Frontiers of the Indus Civilisation. Delhi: Books and Books, pp. 301-10.
-. 1989. Fouilles de Shortughai. 2 Vols. Paris: Diffusion de Boccard.
Jansen, M. 1987. Preliminary Results of ‘Forma Urbis’ Research at Mohenjo-Daro. In M. Jansen and G. Urban (eds.), Interim Reports. Vol. 2. Aachen: ISMEO/RWTH, pp. 9-22.
-. 1989. Water Supply and Sewage Disposal at Mohenjo-Daro. World Archaeology 21(2): 177-92.
Jarrige, J.F. 1989. Excavations at Nausharo 1987-88. Pakistan Archaeology 24: 21-67.
Joshi, J.P. 1990. Excavations at Surkotada. Delhi.
Kenoyer, J.M. 1998. Ancient Cities of the Indus Valley Civilisation. Karachi: Oxford University Press.
Kesarwani, A. 2000-2001. Balu: A Fortified Rural Settlement in Haryana. Puratattva 31: 140-52.

Khan, F.A. 1965. Excavations at Kot Diji. Pakistan Archaeology 2: 11-85.
Lal, B.B. 1979. Kalibangan and the Indus Civilisation. In D.P. Agrawal and D.K. Chakrabarti (eds.), Essays in Indian Protohistory. Delhi: B.R. Publishing, pp. 65-97.

Mackay, E. 1938. Further Excavations at Mohenjo-Daro 2 Vols. Delhi: Archaeological Survey of India.
—_. 1943. Chanbudaro Excavations. New Haven.
Majumdar, N.G. 1934. Explorations in Sind. Calcutta: Archaeological Survey of India.
Marshall, J. (ed.). 1931. Mohenjo-Daro and the Indus Civilisation. 3 Vols. London: Arthur Probsthain.
McMurdo, J. 1834. An Account of the Country of Sindh; with Remarks on the State of Society, the Government, Manners, and Customs of the People. Journal of Royal Asiatic Society 1: 223-57.
Mockler, E. 1877. On Ruins in Makran. Journal of Royal Asiatic Society 9: 121-34.
Mughal, R. et al. 1996. Archaeological Sites and Monuments in Punjab. Preliminary Results of Explorations: 1992-96. Pakistan Archaeology 29: $1-474$.

Nath, A. 1997-98. Rakhigarhi: A Harappan Metropolis in the Saraswati-Drishadvati Divide. Puratattva 28: 39-45.
—. 1998-99. Further Excavations at Rakhigarhi. Puratattva 29: 46-49.
—_. Rakhigarhi 1999-2000. Puratattva 31: 43-46.
Pant, Mohan and Shuji Funo. 2005. The Grid and Modular Measures in the Town Planning of Mohenjo-Daro and Kathmandu Valley. Journal of Asian Architecture and Building Engineering 4 (1): 5159.

Pramanik, Shubhra 2003-04. Excavation at Juni Kuran 200304: A Preliminary Report. Puratattva 34: 45-67.
Rao, L.S. et al. 2003-04. Unearthing Harappan Settlement at Bhirrana (2003-04). Puratattva 34: 20-24.

Rao, S.R. 1963. Excavation at Rangpur and Other Explorations in Gujarat. Ancient India 18-19: 5-207.
—_. 1979. Lothal, a Harappan Port Town. 2 Vols. Delhi: Archaeological Survey of India.
Shinde, V. 2004. Saurashtra and the Sites of Padri and Kuntasi. In D.K.Chakrabarti (ed.), Indus Civilisation Sites in India-New Discoveries. Mumbai: Marg Publications, pp. 62-68.
Snead, R.E. 1967. Recent Morphological Changes along the Coast of West Pakistan. Annals of the Association of American Geographers 57: 550-65.
Stein, A. 1929. An Archaeological Tour in Waziristan and North Baluchistan. Calcutta: Archaeological Survey of India.
—_. 1931. An Archaeological Tour in Gedrosia. Calcutta: Archaeological Survey of India.
Trivedi, P.K. and Patnaik, J.K. 2003-04. Tarkhanewala Dera and Chak 86 (2003-04). Puratattva 34: 30-34.
Vats, M.S. 1940. Excavations at Harappa. Delhi: Archaeological Survey of India.
Wheeler, R.E.M. 1947. Harappa 1946: The Defences and Cemetery R 37. Ancient India 3: 58-130.
[DKC]

Editorial Postscript: Some Recent Researches

The first development, after the volume was in the final stages of publication, is the find of an antiquity without context; it can be stylistically located in north Baluchistan and throws unexpected light on some unknown rituals immediately before or immediately after the beginning of the Indus civilization. The find has been discussed in detail by M.Vidale (see his article with K. Bhan in this volume) in a brief monograph entitled The lady of the spiked throne. The power of a lost ritual (Trieste 2012).

This is an approximately 57 cm long and 22 cm broad terracotta boat-shaped object which has its front portion in the shape of a horned and humped bull’s face. If one looks at the raised and bent legs of the bovine at the sides, this may also give the impression of a chariot. The ‘boat’ has an arched, wagon-like cover at the back, with a seated female figure having a small bull figure on either side as a kind of hand-rest and with a small table to put her feet on. She sits on a throne decked with seven spikes at the back. As Vidale describes,

Fig. 42.

At rear, astern, the boat has vault-like roofing, very similar to the cloth coverings of the wagons of the 19th century … It shelters the figurine of an enthroned lady who is the most important personage of the model … The roofing ends with a kind of vertical handle, painted black, heavily darkened and rather worn…
suggesting that the object underwent a long or intensive manipulation, being grabbed and lifted from this handle, before the presumed final burial.
Below the vault, the dominating female figure, larger in size than the rest of the figurines, sits on a throne, the feet resting on a low cubic stool. The back of the throne has seven spikes, five vertical on top and two horizontal at the sides…; the armrests are delicately shaped as standing humped bulls… Beneath the same vault, the enthroned lady is surrounded by four male ‘attendants’ standing
Against the inner walls … a red-slipped staircase leads to the rear of the forecastle.
The central space of the model, between the roofed throne and its occupant and the prow, is occupied by a double row of figurines sitting on as many low square stools, separated by a central aisle… The figurines are arranged in alternating female and male couples, females coming first and being slightly larger than males .
The object apparently stands for a group ritual, and the way the ‘goddess’ observes from the back a succession of male and female figures on either side of a
central aisle in front implies a highly conceptualised form of a ritual.
h
The second is the publication of an article -L.Giosan et al. (2012), Fluvial Landscapes of the Harappan Civilization, Proceedings of the National Academy of Sciences of the United States of America, Environmental Sciences
The basic contention of the paper is the following:

• aridification intensified in the Harappan territory after $c .5000 \mathrm{BP}(c .3000 \mathrm{BC})$
• the river flows slowed down, leading to gradually decreased flood intensity
• the gradual decrease in flood intensity led to intensive agriculture, leading in turn to urbanisation
• there was a further decline in monsoon precipitation, leading to the weakening of the rivers dependent on the monsoon rains, increasing “the vulnerability of agricultural production supporting Harappan urbanism, leading to settlement downsizing, diversification of crops, and a drastic increase in settlements in the moister monsoon regions of the upper Punjab, Haryana and Uttar Pradesh”.

it is easily noted that this premise of a more or less continuous aridity with more intensification after some time tallies more or less with the premises of a number of scholars who published earlier on the same theme of aridity. From this point of view, it is not a novel idea.