Vivien de Saint-Martin showed a geographer's insight when he remarked, 'The ancient

designation of Sarasvatī very much appears to have embracedm apart from the chief watercourse
flowing far to the west, the totality of the streams flowing down from the mountain close to each
other before they unite in a single bed.1 In other words, he regarded all the streams we have seen
so far- from west to east, the Ghaggar, the Markanda (and we should add the Dangri or Tangri
between these two), the Sarsuti and the Chautang, and their smaller tributaries- as being,
collectively, the relic of the Rig Veda’s Sarasvatī. (The Chautang, as we will see, was later identified
with the Vedic Drishadvatī river).
Vivien de Saint-Martin added the Veda’s description of the Sarasvatī as a river ‘flowing to
the sea’, which, to him, indicated that ‘its course then extending between the Satlej and the gulf
of Kotch. (The Rann of Kachchh- oder spellings include Kach, Kutch, etc.) is a vast marshy and salty
expanse in Gujarat, north of the island of Kachchh and south of the Indo-Pakistan border  see
maps in Figs 1.5, 1.7 and 4.2)
str. 21.
The study of the region fully confirms the Vedic piece of information. The trace of the ancient
riverbed was recently found, still quite recognizable, and was followed far to the west2, a reference
to the explorations of Tod, Colvin and Mackeson. Vivien de Saint-Martin summarized their
findings, which, to him, ‘confirmed the correctness of the tradition’. 3 His own conclusion was, a
full understanding of Vedic geography4, a clue later scholars made ample use of.
He was probably the first scholar to spell out the problem in such clear terms: the Rig Veda refers
to a mighty river calles Sarasvatī, and the topography of the Yamuna-Sutlej interfluve is scarred by
a now dry river system, one of the streams of which still bears the name of Sarsuti (Sarasvatī). Can
the two be equated? His answer in the affirmative was accepted by generations of Sanskritists and
Indologists, some of whom we will talk about in the next chapter.

PUNJAB’S ‘SACRED RIVER’

The same answer sonn made its way into the gazetteers published by the colonial powers. In 1885,
under the entry ‘Ghaggar’, the encyclopedic Imperial Gazetteer of India described the river’s cours,
and noted: ‘In ancient times the lower portion of the river seems to have borne the name of its
confluent the Saraswati or Sarsuti, which joins the main stream in Patiala territory. It then
possessed the dimensions of an important channel… At present, however, every village through
which the stream passes diverted a portion of its waters for irrigation, no less than 10,000 acres
being supplied from this spurce in Ambala District alone… During the lower portion of its cours, in
Sirsa District, the bed of the Ghaggar is dry from November to June, affording (dajući, dopuštajući)
a cultivable surface for rich crops of rice and wheat.’5

1
Vivien de Saint-Martin, Louis Étude sur la géographie et les populations primitives du nord-ouest de l'Inde, op.
cit., p. 20.
2
Vivien de Saint-Martin, Louis Étude sur la géographie et les populations primitives du nord-ouest de l'Inde, op.
cit., p. 22.

3
Ibid., p. 23
4
Ibid, p. 24
5
Hunter, W.W., 'Ganjam' to 'India', Imperial Gazetteer of India, vol. 5, Trübner&Co., London, sec.edn, 1885, pp.
54-55.
Let us turn to the entry ‘Saraswati (Sarsuti)’, defined as a ‘sacred river of the Punjab, famous in
the early Brahmanical annals’. We learn that the river rises ‘in the low hills of Sirmur State,
emerges upon the plain at Zadh Budri (Ad Badri), a place esteemed sacred by all Hindus’, and,
before joining the Ghaggar, ‘passes by the holy town of Thanesar and the numerous shrines of the
Kuruksetra, a tract (‘Kuruksetra’ was originally the name of a region, not just a town as it is today)
celebrated as a centre of pilgrimes, and as the scene of the battle-fields of the Mahabharata’. The
Gazetteer (rječnik geografskih imena) repeats, ‘In ancient times, the united stream below the
point of junction appears to have borne the name of Sarsuti, and, undiminished by irrigation near
the hills, to have flowed across the Rajputana plains…’
Correlating geography with early literature, the Gazetteer adds, ‘Some of the earlest Aryan
settlements in India were on the banks of the Sarasvati, and the surrounding country has from
almost Vedic times been held in high veneration. The Hindus identify the river with Saraswati, the
Sanskrit Goddess of Speech and Learning’.6
We will come to the reasons for this identification in due time; for the moment, we must
turn to fresh explorations of the region’s topographical features, which sought to pinpoint more
precisely how the river came to be ‘lost’.

CHANGING COURSE
Richard Dixon Oldham, a British geologist, joined the Geological Survey of India (GSI) in 1879 at
the age of twenty-one; it ran in the family: his father, Thomas Oldham had been the GSI’s first
director. But scientific posterity remembers the son more than the father: apart from reference
works, memoirs and numerous research papers on India’ geology, R.D. Oldham specialized in
seismology; when a terrible earthquake struck Assam in 1897, destroying Shillong (which was then
part of that state), his study of the seismographic records led him to deduce the existence of the
earth’s molten core. Ill-heallth forced Oldham to leave the GSI and India at the age of forty-five,
though he continued to contribute to the discipline from his retreat in England and later southern
France.
It is his lesser-known research that concerns us here: his upstream (rather, ‘upbed’) survey
from the Bahawalpur region to the Hissar district, in his capacity as deputy superintendent of the
Survey. Understandably, his professional competence gave him an edge over his military
predecessors.
Writing in 1886 in the Journal of Asiatic Society of Bengal7 (Fig. 1.4), he rejected theories
of the day that attributed the loss of the Sarasvatī to diminished rainfall, pointing out that this
would have affected all rivers equally. Instead, he proposed that the ‘Lost River of the Indian
Desert was none other than the Sutlej, and that it was “lost” when the river turned westwards to
join the Bias (Beas)’8 This explanation has been broadly endorsed since then, especially in view of
the Sutlej’s sharp west-ward bend near Rupar (or Ropar, or Rupnagar, in India’s Punjab, not far

6
'Ratlam' to 'Sirmur', Imperial Gazetteer of India, vol. 12, Trübner&Co., London, sec. edn., 1887, pp. 261-62.
7
Oldham, R. D., 'On Probable Changes in the Geography of the Punjab and Its Rivers: An Historico-Geographical
Study', Journal of Asiatic Society of Bengal, vol. 55, 1886, pp. 322-43. (Oldham's paper is partly reproduced in
Vedic Sarasvatī, pp- 81-88.)
8
Ibid., p. 340.
from Chandigarh), and the existence of a paleobed that connected it long ago with the Ghaggar,
which flows hardly 50 km away to the east.
Oldham also believed that part of the Yamunā’s waters might have flowed into the
Ghaggar-Hakra bed in Vedic times: ‘It may have been…that the Jumna (Yamnunā), after leaving
the hills, divided its waters… and that the portion which flowed to the Punjab was known as the
Saraswati while that which joined the Ganges was called the Yamuna.’9 In his opinion, that double
desertion of the Sarasvatī, by the Sutlej and the Yamunā, which brought about ‘a considerable
change in the hydrography of the region’, 10 was caused by the well-known waywardness of north
India’s rivers. As a more recent example, he cited the case of the Brahmaputra changing its cours
in the early nineteenth century, a little upstream of its confluence with the Ganges (in today’s
Bangladesh). That waywardness (svojevolja) is nothing but the effect of the very flat flood plains
of the entire Indo-Gangetic basin: here, the phenomena of erosion and sedimentation, which
would hardly be noticeable in the Deccan’rivers, get greatly amplified and trigger frequent shifts
in the watercourses.
In fact, Oldham might have just as aptly quoted Strabo, a Greek savant of the first century
BCE whose Geography of the ancient world remained unsurpassed for centuries. For his
description of northwest India, Strabo relied on the work of several Greek historians, among them
Aristobulus who had accompanied Alexander the Great on his campaign to India. Strabo noted,
for instance:
He (Aristobulus) says that when he was sent on some business, he saw a tract of land deserted
which contained more than a thousand cities with their villages, for the Indus, having forsaken Its
proper channel, turned itself into another on the left much deeper, into which it burst like a
cataract, so that it no longer watered the country on the right, from which it receded, for this had
been raised by the inundations not only above the level of the new channel but even above that
of the new inundations.11
What could have uplifted this strip of land on the right bank of the Indus? Strabo shrewdly
assumed that ‘India is liable to earthquakes is, of course, fanciful, Strabo’s assumption that northwest
India’s seismic activity might cause changes in the course of major rivers was surprisingly prescient; in
fact, this phenomenon has been invoked by geologists in recent years in the context of the Sarasvatī
(and witnessed in the case of the Indus).
At this point, it is worth stressing that the three currently minor rivers that we have focused
on- the Ghaggar, the Sarsuti and the Chautang, along with their tributaries such as the Dangri or the
Markanda- flow down from the Shivaliks in the strip of land between Chandigarh and Yamunanagar,
which is hardly more than 80 km in breadth. A few more kilometres to the west, we find the Sutlej
flowing toward the Indus and the Arabian Sea; and a litlle to the east, the Yamuna winding her
way to the Ganges and the Bay of Bengal. In other words, these seasonal rivers are located on a
narrow and fairly flat (The average gradient is roughly 300 m over 1000 km, that is, an
imperceptible 30 cm/km or 0.03 per cent- 300 m is the average altitude of the plain below the
Shivalik Hills, and 1000 km the approximate distance to the Rann of Kachchh) watershed between

9
Ibid., p. 342
10
Ibid., p.341
11
Strabo, Geography, book XV, I.19, tr. John W. McCrindle, Ancient India as Described in Classical Literature,
1901; reprinted Oriental Books Reprint Corporation, New Delhi, 1979, p. 25.
the two vast river systems of the 3000 km- long Indus and the 2500 km- long Ganges: It is easy to
visualize how a slight uplift of the sort reported by Strabo, or else erosion caused by poweful
spates, could have triggered the diversion of the Sutlej or Yamun waters westward and eastward,
respectively, away from the Ghaggar-Hakra system, a diversion cogently explained by Oldham for
the first time.
str. 26
That explanation apart, R.D. Oldham’s work was valuable because it established that the
landscape of today’s Punjab and Haryana must have been radically different at some remote
time.

THE RAIN
Five years later, a monumental paper on the hydrography of the Iindus basin appeared under
the pen of Henry George Raverty, a British major of the Indian Army who had acuired first-
hand knowledge of the Punjab during the military campaigns that culminated in the
subjugation of the state. But there was a schoolarly side to this army officer, as he authored
in 1849 a gazetteer of Peshawar, perhaps India’s first district gazetteer. He also learned the
Pashto language of Afghanistan so well that he went on to author a grammar book and a
dictionary of it, and translated into English selections from Afghan poetry.
Our poetic major’s paper, published in 1892, dealt not with the Sarasvatī, but with the
identity of the ‘Mihrān of Sind’, a river reported by the eighth-century Arab invaders to be
flowing east of the Indus in a course parallel to it.12 We need not go into Raverty’s intricate
analysis of historical evidence and discussion of the relative shifts in the beds and confluences
of this river system, especially as many important details remain disputed.13
42 44
43 45
It will be enough for our purpose to highlight some of his conclusions: ‘Sursuti (Sarsuti) is the
name of a river, the ancient Saraswatī… Sutlaj (Sutlej) was a tributary of the Hakrā or Wahindāh, which
was nothing but the bygone Mihrān, and flowed down to the vast salty expanse of the Rann of Kachchh
through the Eastern Nara. The Nara (see Fig. 1.7), str. 27. now a dry channel, has often been assumed
to have been an ancient outlet of the Indus, whether a perennial or a seasonal one; it splits into two
channels on either side of the Rohri Hills of Sind, and is the eastern branch that would have received
the Hakra’s waters. The Hakra’s drying up, which according to Raverty took place in the fourteenth
century CE, ‘reduced a vast extent of once fruitful country to a howling wilderness, and thus several
flourishing cities and towns became ruined or desert by their inhabitants’.

12
Raverty, H.G., 'The Mihrān of Sind and Its Tributaries: A Geographical and Historical Study', Journal of the
Royal Asiatic Society, vol. 61, no.1 & extra number (1892), pp. 155-206 & 297- 508.
13
E.g. Wilhelm, Herbert, 'The Shifiting River: Studies in the History of the Indus Valley', Universitas, vol. 10, no,
1, 1967, pp. 53-68.
 Raverty also traced the name ‘Hakra’ to the Sanskrit sāgara or ‘ocean’, an etymology that has
been largely accepted and explains variants such as Sankra and Sankrah, terms used in Islamic
chronicles.
Raverty’s work in the lower reachs of the Hakra was supplemented by that of Robert
Sivewright, an officier from the Public Works Department (PWD), who spent a few months exploring
the Rann of Kachchh and its geological features, trying in the process to reconstruct some of its history
from the days of Alexander’s campaign in the Indus Valley to the Arab conquest of Sind in the eighth
century and beyond. In 1907, having retired and returned to Britain, Sivewright presented a paper on
‘Cutch and the Ran’ at the Royal Geographical Society, a lecture which R.D. Oldham also attended.14
(44 46)
One of Sivewright’s important observations on this forbidding region was the ‘silting up of the
Greater Ran by the Hakra’,15 in his opinion, this silting caused the gradual build-up of the Rann, which,
he observed, was still navigable even for some time after the Arab conquest.16 But a tectonic uplift of
the region may well have been a contributory factor to the drying up of the Rann, as a 1967 study of
the uplift of the nearby Makran coast suggests.17
Be that is it may, Sivewright’s conclusion that ‘the Ran is the delta of the Hakra, the lost river
of Sind’18 is of greater relevance to us. Sivewright thus agrred with Raverty that the Hakra river kept its
name all the way to its estuary north of the Rann, and his map (Fig. 1.5) makes this even more explicit.
‘RUINS EVERYWHERE’ str. 29.
In his 1887 paper, R. D. Oldham had frequently referred to an article published anonymously in the
Calcutta Review (Fig. 1.6) thirteen years earlier,19 from which he borrowed the idea that the Hakra-
Nara had been the Sutlej’s original bed. The anonymous paper’s author, Oldham informed us,20 was a
namesake (though probably no relative): C.F. Oldham, a surgeon-major in the Indian Army. The
surgeon-major distinguished himself by medical notes (on malaria in particular), but, like Raverty, is a
remembered for his varied scholarly interests: among them, the origins of serpent worship in ancient
cultures, inluding India’s, and the Sarasvatī river.

14
Sivewright , Robert, ‘Cutch and the Ran’ at the Royal Geographical Society, vol. 29, no. 5 May 1907, pp. 518-
35. I am indebted to Prof. R.N. Iyengar for drawing my attention to Sivewright's paper, and to
15
Sivewright, Robert, 'Cutch and the Ran', op. cit., p. 528.
16
(48) Ibid., p. 530
17
(49) Snead, Rodman E., 'Recent Morphological Changes along the Coast of West Pakistan', Annals of the
Association of American Geographers, vol. 57, no. 3, September 1967, pp. 550-65. My thanks to Prof. R. N.
Iyengar for drawing my attention to this paper.
18
(50) Sivewright, Robert 'Cutch and the Ran', op. cit., p. 532.
19
(51) Anonymous (Oldham, C.F.), 'Notes on the Lost River in Indian Dessert', Calcutta Review, vol. 59, 1874,
pp. 1-27
20
(52) Oldham, R.D., 'On Probable Changes in the Geography of the Punjab and Its Rivers', op. cit., p. 322.
 In 1893, almost two decades after his anonymous article, C.F. Oldham examined the whole
issue afresh in a comprehensive and erudite paper entitled ‘The Saraswati and the Lost River of the
Indian Desert’,21 which included a detailed map (Fig. 1.7). He also started off with mentions of the
Sarasvatī in the Rig Veda, and noted that one of its hymns clearly places the river ‘between the Yamuna
and the Satudri (Sutlej) which is its present position.’22 After a brief description of the present course
of the Sarasvatī, Oldham stressed that even after its confluence with the Ghaggar, ‘it was formerly
(known as) the Sarasvatī; that name is still known amongst people…’23 He had no doubt, therefore,
that the lost Rig Vedic river flowes in the bed of today’s Ghaggar. And he added valuable details:
Its ancient course is contiguous with the dry bed a great river which,
as local legends assert, once flowed through the desert to the sea.
In confirmation of these traditions, the channel referred to, which is called Hakra
or Sotra, can be traced through the Bikanir and Bhawulpur (Bhawalapur) States
into Sind, and thence (odande) onwards (odmakao) to the Rann of Katch.
The existence of this river at no very remote period, and the thruth of the legends
which assert the ancient fertility of the lands through which it flowed, are
attested by the ruins which everywhere overspread what is now an arid sandy waste.
str. 32. Throughout this tract are scattered mounds, marking the sites of cities and towns.
And there are strongholds still remaining, in a very decayed state, which were places of importance
at the time of the early Mahommedan invasions.
Amongst these ruins are found not only the huge bricks used by the Hindus in the remote past,
but others of a much later make
All this seems to show that the country must have been fertile for a long period…
Freshwater shells, exactly similar to those now seen in the Panjab rivers, are to
be found in this old riverbed and upon its banks.24  DENI TO NA POČETAK DIPLOMSKOG
HAKRA= WAHIND
NARRA= HAKRA
Pokrajine: Pokram, Bikaner, Nagore, Jaipur, Umarkot, Kotri,

STR. 155

21
(53) Oldham, C.F., 'The Sarasvatī and the Lost River of the Indian Desert', Journal of the Royal Asiatic Society,
vol. 34, pp. 49-76
22
(54) Ibid., p. 51.
23
(55) Ibid. Oldham invokes the Punjab Gazetter of Hissar as evidence (which I have not been able to consult).
24
(57) Oldham, C.F. , 'The Sarasvatī and the Lost River of the Indian Desert', op. cit., p. 54.
New Horizons

A brief visit to four important sites on the Indian side of the border will complement our
acquaintance with Indus-Sarasvatī urbanism: they all bear the Harappan stamp, but because they
sharply differ from each other, they open new horizons on what was already a rich and complex
civilization. Let us start from the Sarasvatī’s upper reaches and go with the current.

BANAWALI

This Harappan city of about 10ha was found in the Fatehabad district of Harayana, on the bank of an
old bed of the Ghaggar. According to R. S. Bisht, who directed the excavations in the 1970s,
‘Banawali’ was an important administrative headquarters or provincial capital and a prosperous
trading centre along the Sarasvatī during the Indus times.’25
The site was already occupied in the pre-urban phase, with some evidence of fortifications and bricks
following the typical Early Harappan proportions of 1:2:3. At the start of the urban phase, ‘all the pre-
existing residential houses were razed to the ground (sravnjene na zemlju) and fresh ones were raised
(uzdignute, izgrađene) with the newly introduced bricks and with thicker walls of better
workmanship’.26 Those ‘newly introduced bricks’ followed the standardized proportions of 1:2:4, and
this ‘razing to the ground’ is one more illustration of the ‘clean-slate strategy (škriljevac- vrsta stijene,
nastaje iz najfinijeg glinenog blata, vrsta crijepa, skup građevinski materijal, znatno postojaniji na kisele
kiše i ostale atmosferske utjecaje prošlost, lisnac)’ we saw in the Indus region.
The site of the Mature phase has a layout (tlocrt, raspored) not found anywhere else so far
(Fig. 7.1), with an overall trapezoidal shape and a semielliptical acropolis. Another unique feature of
Banawali is the presence of a six-metre-wide, V-shaped moat (rov) outside the town’s fortifications,
which was most likely a protection from floods when the river was in spate.
While the streets of the acropolis are mostly at 90o angles, those of the lower town follow a
more complex radial pattern; but several of them are precisely oriented along the north-south axis,
str. 157. and the larger ones are a comfortable 5.4 m wide. Some rich traders lived there, judging from
the presence of seals, hoards of jewellery and stone weights in some of the bigger houses. One of them
boasted a paved living room and a bathroom complete with a raised washbasin!m
Perhaps the most remarkable structure unearthed at Banawali’s acropolis (Fig. 7.2) is a small
building shaped as a semi-ellipse- precisely the shape of the acropolis. As if to make it amply clear that
this was a conscious choice and not an accident, the building harbours an altar that once again

25
Bisht, R.S., 'Excavations at Banawali: 1974-77', in Possehl, Gregory L., (ed.), Harappan Civilization: A Recent
Perspective, op. cit., p. 120.
26
(New Horizons: 2) Bisht, R. S., 'Dholavira and Banawali: Two Different Paradigms of the Harappan Urbis
Forma', Puratattva, no. 29, 1998-99, p. 16.
conforms to a semi-elliptical (or apsidal) shape. There can be little doubt that this building was a small
temple dedicated to fire worship (we will return in Chapter 10).

KALIBANGAN

Some 200 km downstream from Banawali, we come to Kalibangan, on the left bank of the Ghaggar.
Indeed, just a few kilometres further downstream is the confluence with the Chautang, still so
conspicuous on satellite photographs (Fig. 3.3.). Gregory Possehl pust it this way, ‘Kalibangan… is
strategically located at the confluence of the Sarasvatī and Drishadvatī Rivers and must have played a
major role as a way station and monitor of the overland communications of the Harappan peoples.’27

Kalibangan, like Banawali, saw an Early phase, complete with fortifications, rectangular
houses, streets and even drains. In its Mature phase, however, this town embodied a very different
concept of town planning from Banawali’s, even though it was of about the same size: 11.5 ha for the
area within fortifications, and probably a few more hectares outside. Here, the acropolis and the lower
town were twin enclosures, in the form of two oblique parallelograms whose longer sifes were
oriented north-south (Fig. 7.3). In this, Kalibangan followed the general scheme of Mohenjo-daro (Fig.
5.1), whose acropolis occupied a separate mound to the west of the lower town. Mohenjo-daro’s
accropolis is thought to have measured some 200x 400 m, while Kalibangan’s was precisely 120 x 240
m- in both cases, the ratio of length to breadth is 2:1. A massive east-west wall further divides the
accropolis into two rhombs of nearly 120- 120 m each. But Kalibangan is luckier than Mohenjo-daro in
that the lower town’s fortifications are largely traceable, measuring at least 360 x 240 m.
The lower town’s streets formed a well- planned and carefully maintained grid; their widths, starting
from the narrowest, were 1.8 m, 3.6 m, 5.4 m and 7.2 m, in a perfect geometric progression of 1:2:3:4.
(This pattern is partly visible at other sites: we just saw, for instance, strret widths of 5.4 m at Banawali).
As with the overall town plan, we must note the Harappan engineers’ and planners’ foudness for
precise proportions: they did not believe in leaving things to chance, as our ‘modern’ municipal
authorities seemingly do. No urban jungle in thoose protohistoric times!
The only structures permitted on the streets were small brick platforms jutting out near house
entrances, where people evidently sat together (sjedili zajedno) in the evening to chat and exchange
the day’s news: perhaps the arrival of a caravan of traders from Harappa, less than 200 km away, or
the latest gossip from Rakhigarhi and other large urban centres upstream- unless it was simply the
recent harvest in the fields around the town. Houses were, as elsewhere, organized around a central
courtyard, and surplus of wealth (what we call ‘luxury’) is visible in some of them in the form of tiled
floors decorated with the typical Harappan motif of ‘interesecting circles’ (Fig. 7.4).

27
(3) Possehl, Gregory, The Indus Civilizationo, p. 77.
 At Harappa (Fig. 5.3.) the acropolis (on mound AB) is also in the shape of a parallelogram,
measuring roughly 200 x 400 m (the same size as at Mohenjo-daro), while a recessed entracne on its
northern side faces a now dry riverbed of the Ravi. We find a similar device at the northern end of
Kalibangan’s acropolis, facing the Sarasvatī’s dry bed and wide enough to allow carts in and out. str.
160. Such a layout makes eminent sense with rivers acting as important linke between towns and
regions; the recess mush have been designed to afford a measure of control on the movement of
people and goods.
While the northern portion of the acropolis was residential in nature, the southern brought to
light a series of massive brick platforms oriented along cardinal directions. According to B. B. Lal, two
conducted the excavations with B.K. Thapar and J.P. Joshi, the area must have been reserved for ritual
purposes.28 There are several clues to this effect. First, as far as can be judged, it had no regular houses
or other buildings. Second, both accesses to it, through the partitioning wall in the north and an
entrance in the southern fortification wall, were stairways, therefore disallowing the movement of
carts: there must have been a specific reason to compel inhabitants to reach the area on foot. Third
and more explicit, no one of the platform, a row (red, niz) of seven oval-shaped structures, five of them
fairly intact, were found next to each other, sunk in the ground, with a slender stele of clay standing in
the middle of each of them. They contained ash and charcoal, which prompted the excavators to
identitfy them as fire altars. Their location alongside a wall made the officiants sit facing east, the
str. 161. direction still favoured today in rituals; behind them was a half-buried terracotta jar
containing more ash and charcoal; nearby, a few bathing pavements and a well suggest ablutions. In
every detail, the complex, is evocative of religious rituals, and we will return to it when we discuss
Harappan religion. Interestingly, the same kind of altar with a central stele was found in many
individual houses, and Lal attributes a religious purpose to them, since cooking was done in the open
courtyards.
On another of the brick platforms, a carefully built rectangular pit of burnt bricks measuring
1.5 x 1 m contained antlers and bones of bovids, evidently sacrificed as part of a ritual.
LOTHAL
This important site of Gujarat is located some 70 km southwest of Ahmedabad, near the Bhogavo, a
tributary of the Sabarmati river; the Sabarmati flows into the northern end of the Gulf of Khambat (or
Cambay) some 23 km downstream.29 At 7 ha, Lothal is modest in size, though, as often, there is
evidence of habitations extending outside the fortified area (Fig. 7.5). The town’s peripheral wall is

28
(4) This is case of D.R. Bhandarkar, who visited Mohenjo-daro in 1911, ten years before R. D. Banerji, and
concluded that the site was just 200 years old on account of its bricks of 'modern type and not large dimension
like the old'! See Gregory Possehl, Indus Age: The Beginningso, pp. 63-64.
29
(5) Marshall, John, 'Mohenjo-daro', Illustrated London News, 27 February 1926, quoted by McIntosh, Jane R.
A Peaceful Realmo, p. 21.
massive, from 12 to 21 m thick, and was clearly intended to offer a measure of protection against
floods, whose repeated onslaughts left tell-tale marks of ravage on the town and probably brought
about its end.
Lothal’s town planning follows the pattern of Banawali in one respect: the acropolis is within
the town, not separate from it like Kalibangan’s; tucked in the southeast corner, it is demarcated not
by internal fortifications, but buy a separate platform of mud bricks almost four metres high. It has
wide streets, well-designed drains, and a row of twelve bathing platforms in a perfectly straight lin (Fig.
5.4 shows a few of them)- a layout that hints at more than a purely utilitarian purpose. The acropolis
also boasts a large building identified as a warehouse, with square platforms where we can visualize
the goods being packed, tied, sealed, lifted on the shoulders of coolies and finnaly taken away. The
building seems to have suffered a fire, as some of the mud bricks are partly burned; but we should be
grateful for that, as otherwise the sixty-five sealing (impressions of seals on clay) found there might
not have been preserved; some of those sealings still bore the impression of the ropes tied around the
bundles of goods waiting to be shipped.
The lower town reveals considerable industrial activity dealing with beads of various semi-
precious stones, shells and metal working among the other crafts, all of them using techiniques that
have been well-documented at other Harappan sites. The presence of sacrificial and fire altars recalls
the structures found at Kalibangan, but here they are found in individual houses or in streets. We will
study one of them in Chapiter 10.
Fig. 7.5. Ona lepa slika od Lothala: An artist’s impression of Lothal (Gujarat). Note the dockyard in the
foreground and the warehouse on the left. The lower town is in the background. (Not to scale.)
str. 163. The most remarkable structure, alongside the town’s eastern side, is a 217-m-long, 36-m-wide
basin. (Fig 7.6). (Its proportions, incidentally, are almost exactly in the ratio of 6:1.) If were consider
that its 1.5 to 1.8 m-thick walls were made of millions of carefully adjusted baked bricks, we will have
an idea of the energy and resources deployed on its construction. No other Harappan site has so far
come up with such a huge water structure (as long as two-and-a-half football fields! nog. igralište). In
view of stone anchors and marine shells found in it, S.R. Rao, the excavator, identified it as a tidal
dockyard (plovidbeno pristanište): at high tide, boats sailing up the Gulf of Cambay would have easily
pushed on upstream the Bhogavo befor berthing at Lothal’s basin. There are other considerations, too:
its vertical walls are ideal for such a purpose (svrha, namjena), and the flat top of the town’s eastern
fortification would have acted as a wharf; the proximity of the warehouse with its numerous sealings
is another clue to the export of goods; a seal evocative of contact with the Persian Gulf was found
elsewhere at Lothal; and an inlet channel was identified at the basin’s northern end, while a spillway
for overflowing waters was spotted at the southern end.
 Here again, not everyone agreed that it was a dockyard, especially in view of two sharp bends
in the defunct stream leading to the dockyard. But the alternate theory of a water reservoir has
Slika 7.6 Lothal’s huge baked-brick basin, almost 220 m long, thought to have been a dockyard.
str. 164. been rather less convincing: the basin would have been unconscionably large for the purpose;
a normal reservoir would be expected to have slanted walls with steps leading down into it (as we can
see at Dholavira); and what is more, sullage frome one of the town’s major drains emptied right into
the basin.30 A recent study, by three Indian scientists, of multi-spectral satellite imagery combined with
an analysis of sediments around Lothal has lent considerable support to the concept of a tidal dock:
the evidence of former estuaried inland demonstrate that the sea level was higher during Lothal’s
heyday, while an analysis of the satellite photographs reveal that’a meandering, tidally influenced river
flowed from the north past Lothal. Tidal waters would thus have been used to approach up to and
slightly beyond the town of Lothal’,31 which is just what S.R. Rao proposed some four decades earlier.
It may, however, be that Lothal was not a point of direct export, and the small river boats
rather than larger seafaring ships took the goods away to a point of trans-shipment on Saurashtra’s
coast. But Lothal’s association with the sea seems clear enough. Even today, at high tide, sea water
enters the lower reaches of the old riverbed. And when the excavators first explored the site, they
found that at that spot villagers were worshipping a seagoddess named Venuvatimātā. But soon, an
incident occurred which aroused the goddess’s displeasure. Let us hear the story from Rao himself:
Before extending the operations to this sector (of the warehouse) the stones in worship representing
the sea-godess had to be removed against the wishes of the labourers. A few days later there was an
accident resulting in injury to some labourers and the death of one of them. Immediately the labourers
attributed the accident to the sacrilege committed by us in removing the goddes from her original
place of worship, and refused to work on the site. They were later satisfied when the goddess was re-
installed elsewhere with some ceremony. This incident is particularly mentioned here to show how
strong is the tradition of worshipping the sea-goddess at Lothal.32
DHOLAVIRA33
Discovered by J.P. Joshi in 1966 on the Khadir island of the Rann of Kachchh, and excavated two
decades later under the direction of R.S. Bisht, this site had quite a few surprises in store for the

30
(6) For a fuller discussion, see Lal, B.B., India 1947-1997o, p.71.
31
(7)Khadkikar, A.S., C. Rajshekhar& K.P.N. Kumaran, 'Palaeogeography around the Harappan Port of Lothal,
Gujarat, Western India', Antiquity vol. 78, 2004, no. 302, p. 901.
32
(8) Rao. S. R., Lothal: A Harappan Port Town, op. cit. 21.

33
(9) Dholavira's data is mostly from three papers by Bisht, R.S.: Dholavira Excavations: 1990-94o, in Joshi, J. P.
(ed.), Facets of Indian Civilization: Essays in Honour of Prof. B. B. Lal, Aryan Books Internation, New Delhi, 1997,
vol. I., pp. 107-120; 'Dholavira and Banawali: Two Different Paradigms of the Harappan Urbis Forma', op. cit.,
pp. 14-37; 'Urban Planning at Dholacvira: a Harappan City', op. cit., pp. 11-23.
archaelogical world, not the least of which was its very location, in the middle of what is today a harsh
and arid landscape. But we saw how the sea reached higher in the Gulf of Khambat, and in a study of
changes in the sea level around Gujarat, U.B. Mathur recently argued that this was also the case in the
Rann: in Mature Harappan times, it was a ‘shallow arm of the sea’,34 and therefore navigable. (Indeed
we know from Greek records that it was still partly so in the first century BCE.35) It is likely, then, that
Dholavira had easy access to the sea.
Slika 7.7. : Plan of Dholavira, in the Rann of Kachchh (C ASI)

str. 166. In the opinion of the Allchins, ‘Dholavira appears to be one of the most exciting discoveries of
the past half century!’36
Exciting for the following three reasons at least. The first is its unique town planning (Fig. 7.7): though
it followed some of the classical Harappan norms, its overall concept departed from everything we
have seen so far. At 47 ha, Dholavira’s fortified area is four times that of tKalibangan. The city is
bracketed between two small seasonal streams. As at Lothal, its acropolis is inside the city, but as at
Kalibangan, it consists of two adjoining fortified enclosures of similar size, named ‘Bailey’ and ‘Castle’
by the excavator, the latter no doubt because of its massive walls of mud bricks flanked by dressed
stones. Remarkably, the bailey’s dimensions, 120 x 120, are exactly those of the two portions of
Kalibangan’s acropolis. Since Lothal’s acropolis (119 x 118 m on an average) is also of almost the same
size, we can rule out the play of chance: Harappan architects had precise norms in mind, just as
craftsmen had with weights or seals, and buiders with brick proportions.
But the similarities end here. While Mohenjo-daro and Kalibangan appear to be based on a
duality between separate upper and lower towns, Dholavira’s plan is essentially triple: just north of
the acropolis lies the middle town, neatly criss-crossed by broad streets at right angles. While the bailey
and the castle must have been home to the city’s rulers and officials, the middle town perhaps
sheltered traders and craftsmen. It boasted a huge ‘stadium’ or ceremonial ground, over 283 m long
and 47.5 m wide, which must have witnessed elaborate public events. (Its proportions, incidentally,
are 6:1, exactly those of Lothal’s basin.) It had four long and narrow terraces on its southern side, which
suggests a provision for seating. Towering over it, an imposing gate through the castel’s northern wall
led down to a ceremonial pathway that descended on to the staduim: we can almost picture rulers

34
(10) Mathur, U.B., 'Chronology of Harappan Port Towns of Gujarat in the Light of Sea Level Changes during
the Holocene', Man and Environment, vol. XXXVII, 2002, no. 2, p. 64. It is doubtful however, that Dholavira was
actually a port town' as proposed by Mathur, as, unlike Lothal, it does not seem to have had berthing facilities.
35
(11) Periplus of the Erythream Sea, see quatation and discussion in Iyengar, R.N.& B. P. Radhakrishna,
'Geographical Location of Vedic Irina in Southern Rajasthan', Journal of the Geological Society of India, vol. 70,
November 2007, pp. 699-705. Also Iyengar, R. N., B. P. Radhakrishna & S. S. Mishra, 'Vedic Iirina and the Rann-
of-Kutch', Puratattva, no. 38, 2008, pp. 170-180.
36
(12) Allchin, Raymond&Bridget, Origins of a Civilizationo, p. 165.
and high official leading the procession, though the nature of the ceremonies and other activities
enacted there can only be guessed.
Beyond the middle town lay the lower town with habitations in its northeastern and eastern
sectors; common workers probably lived there.
str. 167. With architectural hierarchy being a reflection of social stratification, it would be tempting
to interpret the three succesive enclosures and the presence of more habitations outside the city’s
fortifications as the signs of a functional caste system; yet we must resist the temptation until much
more is known about Harappan society and its internal workings. (Let us keep in mind, too, that a
Kshatriya or warrior class is conspicuously absent from Harappan society.)
Dholavira is the only known Harappan site where stone was used on such a scale. Stone
dressing was done with chisels of hardened bronze, and we will have some inkling of the task involved
iw we remember that the castle’s fortifications, up to 18.5 m wide in places, were made of mud bricks
flanked by high stone walls; within its width a few rooms were built with dressed stones, and in some
of them highly polishef segments of pillars, both square and circular, were found in their original places
(Fig. 7.8) Slika: A room in the eastern gate of the castle’s fortifications at Dholavira. Note the two
square of stone pillars and one circular segment. (C ASI)
str. 168. Such segments, with a central hole, were piled up on top of one another, and when the desired
height was reached, a wooden pole was inserted through the whole column to keep it together- an
ingenious alternative to monolithic pillars.
There is an unexpected feature in Dholaviras’s town planning, which has to do the specific
proportions followed by its enclousures, and we will have a peep at it in Chapiter 9.
Water conservation is Dholavira’s second hallmark, inseparable from the first. The city had a
few wells, with the most imposing of them (Fig. 7.9) being in the castle, but great care was taken to
store every possible amount of rainwater: a series of huge reservoirs hugged the castel’s eastern and
southern fortifications; the largest two measured about 73 x 29 m and 33 x 9 m respectively, with the
latter carved out of massive rock, making it, in Bisht’s opinion, ‘the earliest ever rock-cut example’37 of
water structure (Fig. 7. 10). They were partly fed by rainwater harvested from the castle, where
complex stone structures were built to that effect.
Slika 7. 9.: A stone well in Dholavira’s castle. (The woman at the bottom gives the scale.) Note the
grooves on the stone slab, made by sliding ropes. (C ASI)
str. 169.
Elsewhere, huge stone drains, high enough for a man to walk trough, directed storm water to the
western and northwestern sections of the lower town separated by broad bunds, creating in effect as

37
(13) Bisht, R.S., 'Dholavira and Banawali: Two Different Paradigms of the Harappan Urbis Forma'.op.cit.,p. 28.
many reservoirs. Their main supply, however, came from the two seasonal streams to the north and
south of the city, whose waters were slowed down by a series of dams and partly deflected to the
lower town. Altogether, as much as a third of Dholavira’s area was intended to conserve water: in
effect, the monsoon must have turned it into a kind of lake city.
Two important conclusions flow from Dholavira’s skills in water management-which, once
again, will long remain the envy of our modern Indian cities. First, the size of the storm-water drains
points to the sudden inrush of water during heavy rains, while the dams, the rainwater harvesting
structures and the sheer hugeness of the reservoirs reflect a desire to save every drop of the precious
liquid: rains must have been rare overall. In other words, the pattern of rainfall was more or less what
it is today in Kachchh.
Slika 7. 10.: Dholavira: a huge rock-cut reservoir (‘SR30), south of the castle.
str. 170. Second, we find at Dholavira the same obsession with water as at Mohenjo-daro: there, most
archaeologists have seen in it a religious trait, keeping in mind the Great Bath, the vast number of
wells, and the luxury of the bathrooms.
Its town planning apart, Dholavira shot into prominence because of a unique find: an
inscription almost three metres long, found lying on the floor of one of the chambers of the castle’s
northern gate. It was not inscribed there; its ten signs, each over 35 cm high, were made of a crystalline
material which must have been embedded in a wooden plank, and the whole ‘signboard’ was probably
hung above the northern gate, where it would have been visible to musch of the middle town. In terms
of size, there is no remotely comparable inscription from any other Harappan site. (Of course, boards
with signs simply carved or painted on a plank would have vanished without trace; it is the crystalline
material alone that was preserved in this case.)
Although the Dholavira signboard has not helped crack the script, it does show that a
substantial number of people- at least those assembled on the ceremonial ground just below, in this
case- were expected to read it. This gives the lie to earlier theories that knowledge of the script was a
reserved for a small ‘elite’.
After a few centuries, Dholavira’s admirable urban order suffered an eclipse.  POVEŽI S
DUPUIS, 1963, KOLIKO JE TRAJAO VRHUNAC CIVILIZACIJE!!
The civic maintenance came to be neglected; the city shrank to a small settlement, which was
eventually abandoned. After an interval, new people came and occupied the site for a while, but their
rough circular dwellings (boravište, mjesto stanovanja) had no connection with the previous planned
houses, nor were any of the classical Harappan features visible. Gone was the splendour of the city
with its massive acropolis towering over wide streets and huge water reservoirs.
The new dwellers could not stay very long. The last standing buildings crumbled, and sand and
mud slowly buried the ruins, sending them to sleep for some four millennia.