Вы находитесь на странице: 1из 40

:; 18 :;



( A ) »

It has already been pointed out that there are mojin-

tains, and spurs of hills only on the northern,eastern and

western extremities of Bengal* Deserts are conspicuous by

their absence in this province. The most prominent physical

features of Bengal are it‘s rivers.

Rivers of Bengal have cleft it asunder into four

major divisions corresponding to north, west, central and

east Bengal* Each of these divisions has a geographical

entity quite distinct from the others.

The central division is formed by the deltaic por­

tion of the Ganges, we admit that the geographical limits

of this Gangetic delta is not unanimously defined by scholars.

1. Delta is originally a Greek alphabet which in capital

form looks like • In physical geography the name is

applied to a peculiar kind of islands formed by the detri-

tus of rivers at their confluence with the sea or a lake.

The name * delta' was formerly ascribed to th6 fan-shaped

deposit formed by the Nile. Subsequent observations, how­

ever, proved that the Nile is not the only delta-builder

in the world* there are several others also including the

R.IUMookherjee thinks that ’'the whole of Bengal is a

fertile alluvial plain, but this can be divided into four

natural regions according to the extent to which the soil

is enriched by silt deposited when the rivers are in flood.

These regions are (1) the old Delta, or (a) West Bengal and
(b) Central Bengal; (2) the New Delta or East Bengal; and
(3) the Ganges-Brahmaputra Doab or northern Bengal.

On the other hand, in the opinion of Oldham the

"whole country including the Sundarbans proper lying between

the Hooghly on the West and Meghna on the East is only the

delta caused by the deposition of debris carried down by

the rivers Ganges and Brahmaputra and their tributaries."

According to Fox, the deltaic portion in Bengal

begins from the point where the Bhaglrathi and the Padma
bifurcate from each other.

The touchstone for evaluating these rather opiniona­

ted theories is provided by a physiographic survey of Bengal

from the standpoint of geological structure, such a survey,

2. R.Il.Mookherjee, Changing Face of Bengal - a Study in

Riverine Economy, P. 116.
3. Proceedings ASB, 1870, P.47.
4. Fox, Physical Geography for Indian Students, P.189.
20 ;i

however, points out that the districts of B-ankura, Burdwan,

Birbhum and the western half of the Midnapore districts
Belong to the Sub-montane continuation of the Chliotanagpur
plateau. They have only a thin cover of silt resulting
from occasional floods; the surface below is rocky. The
eastern part of the Midnapore district is the “Piedmont
plain built up by the Damodar, Rupnarayan and other streams'1

The region between the Ganges (Padma) (and the

broad Ganges valley above) and the Brahmaputra or the
Ganges-Brahmaputra Doab stretches from the Terai country in
the north to the Ganges (Padma) in the south. Excepting the
mountainous regions in the extreme north, it is alluvial in
character. The land between the Brahmaputra and the Surma is
also alluvial in character. To the east and north-east are

5. K. Bagehi, The Ganges Delta, P. 18.

6. K. Bagehi, op. cit.
7. Ibid. P.19.
8. Ibid
9. Ibid, PP.34r5* In a delta the “Strata are arranged in
a particular fashion bottom set, fore-set and top-set beds;“
(2) a delta ends “with a steep §lope towards the sea,“ but
alluvial plains merge gradually into adjacent countries;“
(3) Channels in the deltas are distributaries or redistribu­
taries, but in alluvial fans they are tributaries.

low hills of Sylhet and Tripura and the Chittagong hill

tracts. Though not of great height, these are very important

landmark, as they define the limits of the possibility of

eastern drift of any Bengal river.

Dr. K. Bagchi has clearly shown that the deltaic.

features are characteristic only of the region hounded by

the Padma on the north, the Bhagirathi on the west and the
Meghna on the east.

Physiographically considered, Bengal thus comprises

the alluvial plains of North Bengal, those between the

Bfeahmaputra and the Meghna, the deltaic region between the

Bhagirathi and the Meghna and sub-montane tracts of the

Chittagong and Burdwan divisions.

10. Op. cit. , P.35.

22 • s

( B )

Bengal's geography thus has been conditioned by
its river systems. The most important of its rivers is
the Ganges.

The Ganges from Hardwar to the Bajmahal hills has

retained its old course which at places is 10/12 miles in
breadth. Below the Rajmahal the Ganges after entering into
the plains of Bengal has bifurcated herself into two main
channels, one running South-east called the Padma and the
other flowing straight southward called the Bhagirathi.

These two channels branch off at Chhabghati near

o o
the police Station of Suti (lat. 24 26‘N. long. 88 2* E.)

in the district of Murshi'dabad. At present the Bhagirathi

flows past Jangipur (Latitude 24. 28 N. 5 Longitude 88.5 E.),
Lalbagh (Lat. 24.13N.; Long.74.1 E.), Katwa (Lat. 23.39 N. j
Long. 88.11 E.), Navadwip and Calcutta and empties itself
into the sea near Diamond Harbour (Lat. 22,11 N.; Long.
88.14 E.)

The Padma carries the main volume of the upland

water of the Ganges. It flows .South-east until it joins the
Meghna near Chandpur and ultimately debouches into the Bay
of Bengal.

11. Kapil Bhattacharyya, Bangla Des'er nad-nadl parikalpana

(in Bengali), P. 10,

Several stories and theories are current/regarding

the origin of these two main branches of the Ganges. Sir
William Willcocks interpretes the story of Bliagiratha as
suggesting that the Padma is the priginal course of the
Ganges and that the Bhagirathi and other rivers of Central
Bengal are canals excavated by 'Hindu', rulers of Bengal
for 'over-flow irrigation'.

Sir Willcocks' view, however, finds no corrobora­

tion from either geological evidence or literature or

According to geologists, in a deltaic plain velo­

city of current is sometimes a deciding factor in fixing
the antiquity of rivers. Secondly, in a loose and soft
soil rivers generally flow through the shortest route to
14 _
the sea. Hydrograpliica 11 y considered the Bhagirathi
channel of the Ganges appears to have formed first.

As early as the second century A. D. Ptolemy,

probably on the evidence of a still earlier source, refers
to the five different mouths of the Ganges, and indicated

12. Ramayana, I, 33 - 34; Brhaddharma - Pur ana, P.22,

# •" ' '' .................. " " " • "

37 - 38 etc.
13. Sir William Willcocks, Lectures on the Ancient System
of Irrigation in Bengal., P.4 ff.
14. K. Bagchi, Op. cit. , P. 40.
15. Ibid. ,
16. Ptolemy, vii, I, 18.
:: 24 ::

great distance between the easternmost and the wefetermost

months. This perhaps suggests the existence of both the
Bhagirathi and the Padma whose mouths were widely*separated

from each other. •

In Sanskrit literature the name Bhagirathi is

applied to the entire course of the Ganges from the Himala-
yas to the sea.

Inscriptions belonging to the eighth and ninth

centuries A.D. refer to the upper course of the Ganges
beyond Bengal as the Bhagirathi. In the Deopara inscription
the Bhagirathi is regarded as a heavenly river and in the
Govindapur copper plate of Laksmanasena the Jahnavi is
described as flowing by Betaddacaturaka (modern Betod in
the district of Howrah).

According to Rennell's map the Bhagirathr flows

past Jungipour, Moorshedabad, Cossimbuzar, Burhampour,
Plassey, Cutwa, Ahgadeep, Nuddeah, Mirzapour, Bansbarya,

17. The longitude and the latitude of the westernmost mouth

o o
of the Ganges of Plolemy are 144 30* and 18 15'
respectively while those of the easternmost mouth are
o o
148 30* and 18 15*.
18. Kumarasambhava , I, 15.
19. A. K. Maitreya, Ganda-lekhamala. P.9 ff.
20. IB. , P.49, V.33.
21. Ibid, pp. 92-98.
: 25 ::

Hooghly Chandernagore, Serampour, Calcutta, Budgebudge, and


An interesting account of the course of the

Ganges in the fifteenth century A/D* is furnished by
Vipradasa (1495 A.D.) in his Manasa-marieala. Here we find

a long description of the journey of the merchant Chand on
his vessel from Rajghat to the confluence of the Ganges with
the sea. The names of the rivers and places through or by
the side of which -dhSnd steered his vessel were as follows :

the river A joy, Uijani, the river Siva, Katwa, the river
Indrani, Indraghata, Nadiya, Phulia, Guptipara, Mirjapura,
Triveri, Saptagrama, Kumarahdta, Hoogly, Bhatpara, Boro,
Kakinara, Mulajora, Gaduliya, Paikpara, Bhadresvara,
Champdani, Icchapura, Bankibajar, Nimaitirtha (present
Vaidyabatl) Chanak, Mahesa, Kha#daha, Sripata, Risira,
Sukachar, Konnagara, Kotarang, Kumarhati, Adiyadaha, Ghusuri,
Chitrapura, Kalikatd, Yetada, Kalighata, ChUda'ghata, Barui-
pur , Chhatrabhoga , Badarikakunda , Hathiagarh, Chaumukhi,

Satamukhi and lastly the confluence of the Ganges with the

22. Proceedings ASB, 1892, pp. 193-97.

23. Ibid., It is interesting to observe that the names of
Kalikata (Calcutta), Hooghly, etc. have been referred to
which were not current in the fifteenth century A.D. It
appears highly probable that the later compilers added
these names to the original list of Yipradasa.
26 ::

The map of Van denBroucke (1660) delineates

many ports and market towns on the BhSglrathi such as
Soti (Suti) , Cassimbazar, Plassi* Neddia, Tripeni (Triveni),
Coatgam (Satgaon) , Oegeli (Hoogly), Bandel, Chandernagor,
Collecatte( Caleutta)etc. Many of thesis names are mentioned
in the account of Vipradasa also. Thus in the seventeenth
century banks of the Bhagirathi abounded in such settle­

The account of Vipradasa (1495 A.D.) and the map

of Van denBroucke (1660 A.D.) do not leave any room for
doubt that the course of the Bhagirathi at least up to
Kalikata (Calcutta) remained unchanged from the fifteenth
to the seventeenth century A.D. But the channel which led
the opening of the Bhagirathi to the sea by the fifteenth
century* as described by Vipradasa, is apparently the same
as the present Adigariga. The map of Van denBroucke also
delineates the Adigaiiga as a considerably wide canal even
though it does not locate any town on its banks, On the
other hand, in the map of Rennel, prepared a century
later, the canal itself cannot to traced at all. Probably
sometime in the seventeenth or the eighteenth century the
bed of the Adigahga began-to deteriorate. And, as its pre­
sent pitiable state indicates, this process of deteriora­
tion was never effectively reversed.

That a channel of the Bhagirathi flowed past

:: 27 ::

Barulpur, Jaynagar-Majilpur, Hathiagarh etc. is evident

from various data. The name Majilapura, literally meaning
•the silted city1, according to a local tradition, perpe-
tuates the memory of dying up of the Ganges in that
locality. Secondly, some large tanks of this region are
* 25
still known by such names as Ghoshrganga, Bose-gahga etc.,
which are probably remnants of the original course of the
Ganges through the area concerned. Again, several old
temples of this locality standing in rows and at places in
two rows having some distance between them, not only indi­
cate that they were originally erected on the banks of
the holy Ganges, but at the same time enable us to chart
out the course of the river up to the Hathiagarh fiscal

Vlpradasa (1495 A.D.) refers to Saptagrama near

which the Sarasvati and the Yamuna branch off from the
Bhagirathi. At present both the Sarasvati and the Yamuna

24. The information was collected by us from the local

people of the region concerned.
25. Hunter, Statistical Account of Bengal. Vol.I, P.29.
26. Ibid.
On the evidence of Copper-plate grants of the Senas
and the Yarmans, N» K.Bhattashali concludes that the
Bhagirathi has not changed it's course at least from
the time of the Yarmans, If the Adi-gariga had ever
been the main o channel of the Ganges, this must
been before C.1025 A.D. (JRAS, 1933, P.85).
27 . Vipradasa, Manas a-Uahgala> p.142
have silted up. In the eighteenth century map of Rennel
they are represented hy narrow lines, the former joining the
Bhaglrathl a little Below Calcutta, In the map of Van den
_ •
Broucke also the Sarasvati, a little wider in extent,
originates from the Ganges near Satgaon only to empty .
itself into the parent stream below Calcutta.

In the sixteenth century map of Jao de Barros

(1550) the course of the Sarasvati is totally different.
Issuing out of the Ganges near Saptagrama, it flows di­
rect South-west and joins the Damodar near the confluence
of the Banka Damodar, referred to by Ksemananda (1640 A.D. )■
in his Manasa-Manga la.

Caesar Frederick (1565 A.D.) writes, “the Saras­

vati above Bator was very shallow and hence only small
ships could go to Sategaon." so already by the middle

28. Ketakadasa Ksemananda, Manasa-Mahgala. p.82.

29. Quoted from J. N.Das Gupta’s 'Bengal in the sixteenth
century, PP. 103 - 4. Caesar Frederick further
writes • " a good tides rowing before you come to
Satagan, you shall have a place which is called
Bullor, and from thence upwards the ships does
not go, because that upwards the river is very
shallow and little water."

of the sixteenth century A.D., this channel had "become

too narrow to accommodate larger ships. Excavation of an

artificial canal between the Hooghly and the SarasvatI

below Calcutta in the time of Nanrab Aliburdi has been

' ‘ 30
accounted for this hydrographic change,

Vipradasa (1495 A»D.) describes Saptagrama as a

- flourishing centre of trade and commerce and the Saras-
vati and the Xumuna as mightly rivers. But curiously

30. Nihar Ranjan Ray, Barielar nad-nadi, P._ 20.

Sarasvati is mentioned in a Muhammedan inscriptions

of the thirteenth century. In the map of Valentijn (based

on information gathered 1660 - 70), it is delineated as a
large stream (O’malley and M,Chakravarty, Bengal District
Gazetteer, 1912, Vol. XXIX, p,112,).

31. Vipradasa, Manasamangala, P. 142,

32. Ibid. Rennel observes, ’’Satgong or Satagong, now an

inconsiderable village on a small creek of the Hooghly
river, about 4 miles to the northwest of Hooghly, was in
1566, and probably later, a large commercial city in which
the European traders had their factories in Bengal.
(Memoir of a map of Hindoostan, P. 57).

enough the vessel of Chand, as described by the poet,

sailed along the Bhagirathi instead of the Sarasvati.
It may indicate that already by the close of the fifteenth
century A.D. the Sarasvati could *not accommodate larger
vessels or was not navigable for larger boats, although,

Saptagrama had not yet lost its importance as an emporium.

As has already been pointed out the map of Jao

de Barros (1550 A.D*) definitely shows that the southward
course of the Damodar was once the channel through which .
the Sarasvati flowed*

It seems reasonable to suppose that still earlier

the Sarasvati branching off the Bhagirathi near Satgaon,
flowed direct south across the Damodar until it joined the
Bupnarayan a little above Tamluk or the ancient port of
Tamralipta. In fact in the Matsya Purana the Ganges

33* D* R. Bhandarkar Volume. P.345* The observations of

Rennel may be cited in this connection. He writes, " I
suspect that its thence course ... was by way of Adaum-
pour, Omptah and Tamlookj and that the river called the
old Ganges was a part of its course, and received that
name, while circumstances of the change was fresh in the
memory of the people. The appearance of the country bet­
ween Satgong and Tamlook, countenances such on opinion .**

(Memoir,' P. 57).

is described as flowing through the countries of the

Brahmottaras, Vangas and Taimraliptas. It appears from
Fa-hsien's account that the Ganga reached the kingdom of
To-mo-li-chi (the capital of) whfch was a sea’-port.
According to K'ang J’ai (of the 3rd century A.D.) as well
Tamralipta was the port where the Hua!ng-chi (Gangi i. e. ,
_ 36
the Ganga) debouched into the sea. Probably by the
name Ganga has been meant the westernmost branch of the
Bhaglrathi which later came to be known as the Sarasvati.
The report of the Stevenson Moore Committee indeed states
that '♦the main branch of the Sarasvati flowed South-west
from Iriveni past Satgaon- and Amta into the Damodar and
probably also the Bupnarayan."

It is interesting to observe that Triveni in

Bengal is not mentioned in the Mahabharata through it
refers to Yudhisthira's visit to the Gahga-Sagara.

34. Mats van urana. 121, 27 ff.

35. Legge, Travels of Fa-Hien. P.100.
36. L. Petech, Northern India according to the
Shui- ching- Chu, P*54.
37. Quoted from D.R.Bhandarkar volume, P. 345. In the
Akbarnama (vol. Ill, P.153) it is stated that "the river
Ganges divides into two branches at Tanda. One goes to the
most of Satgam and ends in Orissa. The other goes to
Mahmudabad, Fatahbad, Sonargaon and Chatgaon."
38. Mbh. , II, P.144.

The Mahapuranas also do not refer to Triveni. The'latter

is mentioned for the first time in the Brhaddharma-Purana.
• •

The Padma-Purana speaks of Tribeni and the junction of the

tliree rivers — Gahga, Yamuna and j;he Sarasvati. In the
Pavanaduta of Dhoyi reference is made to the Barasvati and
the Yamuna emanating from the Bhaglrathi. The fact that*

39. Brhaddharmapurana, I, XXII, 33.

■■ * ' •

40. Padmapurana, 14.

41. Pavanaduta, v, 23, In the Ain-i-Akbari (ed. Jarrett
and revised hy J. M. Sarkar, vol. Ill, P. 133) it is stated
that the Ganges divided into three Streams,' (1) the Sarauti,
(2) the Jamma and (3) the Ganges-called collectively in
the Hindi language Tribeni. The Akbarnama (vol. Ill, p.693)
refers to Trimohena where the Ganges, Jumna and the Sakni
join. Beveridge (Akbarnama, vol. Ill, P.693, n. 5) thinks
Trimohene to be Tribeni in the Hooghly district.

sanctity was attached to Triveni in only late mediaeval

sources may betray that the trifurcation of the rivers at ~

Triveni took place in a comparatively late period.

While commenting on the expression Dskslna-

Pravaga occuring in the Mahabharata, Raghunandana in his
/ 43
Prayascittatattva (16th century)' states that it was to
the South of Pradyumnanagara and Uorth of the Sarasvati.
The topography identifies it with Triveni which was evidetly
considered in the 16th century as sacred as Prayaga of the


This inference suggests that not ail branches and

tributaries of the Bhagirathi can claim great antiquity.

The Bhagirathi itself, however, must be considered as one

of the oldest rivers of India.

The other channel, the Padma has also flowed past

the ages of history.

In the eighteenth century, during the time of

Rennel's survey (1764 - 77) the Padma used to unite with

the Jamuna and with the combined water of the Atreyi and

42. The expression according to Raghnuandana occurs in

the Mahabharata, II, 144.
43* Prayascittatfo.tfcv&>u, Gangamahatmya. P. 100.
44. Ibid.
: : 34 ::

the Karatoya at Juffergung. Thence flowing southward it

joined the Meghna near above the island of DakJhin Shahba-

zpur about twenty five miles below Chandpur.

In the map of Van den Broitcke the main channel

of the Padma flows West of Faridpur and Backergunj. It

seems that by the middle of the 17th century the present

Garai-Madhumati channel was its usual course. A century

earlier in the map of Be Barros, however, the Padma is

shown as flowing past Rampur Boalia, through Chalanjhil,

the Dhalsswari and the Buriganga, past Dacca into the


45. Rennel.f Memoir of a Man of Hindustan. P.345*^.Hamilton

(Geographical. Statistical and Historical Description of

Bindoostan. Vol. I, P. lyy)* observes that the

Ganges once flowed in the tract now occupied by the lakes

and^between Nattore and Jaffergunge.

46. Cf. the map of Van Dan Broucke (1660).

47. In Rennel’s map a branch of the Padma issues out of

the main stream past Gwalpara, Fringyabazar (1st. 23 33'

H. long. 90 23* E) joins the Brahmaputra near Rajabari.

Of Fringyabazar, Hamilton (op.$it; , P.188) writes ’‘this small

town stands on the west side of the Dullasery river formed a
branch of the Ganges and one of the Brahmaputra about 13
miles souths west of Dacca.’1
• • 35 s •

It is apparent from the description of Padmavati,

as given in the account of Tavernier and the Baharistani-
i-Ghyabi by Mirza Nathan, that a channel of the Padma
in the 17th century flowed along €he Dhales^ari or a
parallel course.

The Padma has been referred to by still earlier

51 52
sources. The Mahabhagavata-Purana, the Brahaddharma-Purana,
53 •
the KrttivasI Ramayana
* »
explain the origin of the Padml,

48. Taverniers Travels in India (BK. I). P.102.

49. Baharistan-i- Ghvabi, P. 56 ff.
50. D.R.Bhandarkar Volume, P. 351.
51. Devibhagavatam, IX, 7, 20ff. 5 In the D evib ha gava t am
it is stated that due to the blessings of Kz?sna, Padma
would be known as the Padmavati (ix, 7, 20). The text
further attaches sanctity to the Padma by stating that
the virtuous will take bath in that river (ix, 7, 23-4).
52. Brha ddha rma -. Pur&na. Purvakhanda, 65 Madhyakhanda .
» -- T ■- * *

22, 37 - 38.
53. Ramayana I.
_ _ 54
whereas the Dp.vibhagavatam refers to the origin of the
Ganges and the Padma separately. The Idilpur copper-plate

of Srieandra mentions the Satata-padma vat i-visaya and the

' *
* . t.
Kumara-talaka-mandala. evidently bearing reference to the
• « mmt
Padma and the Kumara rivers respectively.

’ The course through which the Padma carried the

Gangetic water in ancient days is not known to us. We may,

however, refer to the evidence of Ptolemy, who, as. we have

already noted, mentions the five mouths of the Ganges. These


(1) The Kambyson, Poloura, a town.

(2) The Mega.

(3) The third mouth called Kamberikhon,

Tilogrammon, a town.

(4) The fourth mouth called Pseudostomon.

(5) The fifth mouth called Antibole.

Identification of the above rivers cannot be

made with certainly. It is interesting that the difference

between the extreme mouths of the Ganges of Ptolemy is 4

54. See no. n.41.

55. El, XVII, P.190. N.G.Majumdar, however, reads

Satatapadmavati Visaya•
* ... "" r"rrT~ %ni .... • 1 "

56. Ptolemy. Op. cit.

:: 37

of latitude. The distance between Tamluk and Chittagong
is also.the same. Hence it is not unlikely that in

the second century'A.D. the channel, now known as the

Padma flowed through one of the pastern mouths of the


It must, however, he conceded that though the

channel of the Padma existed in the ancient period, the

name itself was not mentioned in the sources datable to

the early centuries of the Christian era. It was also not

universally considered as a channel of the Ganges even

in the early mediaeval period. We have already referred

to a text which distinguished the Ganga from the Padmsu

And, never in the past it achieved that sacred status which

is , associated with the'Bhagirathi.

The reason was, perhaps, that the Bhagirathi of

yore was older, more familiar and, unlike the condition

of the present day, wider than the channel now called the

Pa dma.

57. The Westernmost mouth of the Ganges, according to

o o
Ptolemy is Kambyson (144 13’ , 18 15’) and the east-
o o
ernmost mouth is Antebole (148 30’, 13*15*).

58. The latitude and longitude of Tamluk and Chittagong

o o o o
are 22 18* N, 87 68* E and 22 21* N, 91 53 E

:: 38 ::

No doubt, the Padma is to-day wider and larger

than the other channel and the same has been the case for
a long time.

In the 18th century map of Rennel and also in

the map of Van denBroucke (1660) the. Padma is shown as*
the main channel of the Ganges, whereas the Bhagirathi
appears to be much narrower in extent. Moreover, Major
Rennel designates the Padma by the name of the Ganges.
He writes : "the proper name of this river (the Ganges)
in the language of Hindoostan (or Indostan) is Padda. It
is also called Bara gang or the Great river

This relative position of these branches was

noticeable even in the age of the composition of the
Ramayana by Krttivasa, i.e. , in c. A.D. 1415-16. This
text refers to both tlie Chhotagariga (the smaller Ganges)
and the Bada-ganga (the wider Ganges). It indicates
that in the fifteenth century the Padma was the larger
channel of the Ganges.

59. Rennel, On. cit. , P.255, n.

60. Krttivasi Ramayana, ed. Puma Chandra Dey, P. (66)
Joseph Tieff enthaler (ha Geographic da I^Indoust'an,
tr. byBernaulli, P. (6) refers to the Bhagirathi as the
Ganga minor and the Padma as the Gahga major.
:: 39 ::

In De Barros map (1550) also the Padma is deli­

neated as the wider channel. Abul Fazl and foreign trave­
llers visiting India in the sixteenth century describe
61 *
Padma as the Ganges. Itijh Battuta, a traveller of the
fourteenth century, writes : "the firgt city of Bengali
which we entered was called Sadkawan, a big place on the
shore of the great sea. The Ganges to which the Hindus
go on pilgrimage and the river Jun unite in that neigh-
bourhood before falling into the sea."

Yule has identified Sadkawan with Chittagong

and the Jun with the Jamuna. If the identifications be
correct then it must be maintained that in the fourteenth
century also the Padma was the main channel of the Ganges.

ETo doubt, Dr. R. C. Majumdar observes, on the

basis of the report- of the Stevenson Moore Committee,
that Sadkawan, the Ganges and the Jun of Ibn Battuta may also
be identified with Satgaon, the Bhaglrathl and the

61. Akbar-nama, III, P. 46 ff.

62. Gibb, Travels of Ibn Battuta, P.267.
63. Yule, Cathay and the way thiter : being a Collection
of mediaeval notices of China, Yol.II, P.459, n.1. Yule
thinks that Ibn Battuta actually refers to the junction
of the Ganges (Padma) and the Brahmaputra.
40 • s


We are unable to accept it on two grounds.

First, Ibn Battuta’s Sadkawan was ”on the shore of the

great sea”-—a description which does not tally with

the situation of Satgaon or Saptagrama in the present

Hooghly district. Secondly, the Jamuna near Saptagrama

issues out of the Bhagirathi and does not unite with

the latter.

The existence of the Padma at least as an

important branch can be traced bank to an earlier age.

We are here referring to the Ramacaritam of Sandhyakara-

nandin composed in the 12th century chiefly to narrate

the achievements of the Pala monarch Ramapala who flour­

ished in the same period. It is stated in this text that

Varendri is bounded by the Karatoya on the &ast and the

64. D.R.Bhandarkar Volume. P.348. -

In the Aln-i-Akbari ( Vol. Ill P.133) reference is

made to Triveni,and the third Stream after spreading into

a thousand channels is described as joining the sea at

Satgaon, where the Sarsuti & the Jamuna unite with that


65. Ramacarita , ed. R.G.Basak and R.C.Ma jumdar ,

• * 41 •*

Ganga on the South. It was by crossing the river Ganga
that Ramapala entered into VarendrI, i. e. Worth Bengal.

The topography of Bengal demands that the Ganga of the

Ramacaritam should he identified with the river known at

present as the Padma. This identification and the fact

that the Ganga was a sacred as well asan important river

may indicate that the Padma was already a channel of

considerable importance in the 12th century A.D.

Wo doubt, the evidence of the Rama car it am does

not prove that the Padma was wider than the Bhagirathf

in the 12th century A.D. Wevertheless it'goes against

the theory of Dr. R. C.Majumdar that in the early media­

eval period of the history of Bengal the Padma was not

a river of any considerable importance.

66. Rama carita, III, 10.

67. Ibid. II, 11.

68. Dr. W. R. Ray thinks that the name Padma is traceable

in the compositions of Siddhacarya Bhusuku who lived in

the middle of the eleventh'century A.D. (Banalar-

nad-nadi. PP. 19 - 20).

69. D.R.Bhandarkar Volume. P.346.

42 ::

We admit that the evidence at our disposal is

not sufficient to allow us to fix the date of the intro­

duction of the name Padntfa or to determine the period which

initiated the connection of the .Padma with the appelation

Ganga or that from which the Padma began to he considered

as the wider channel;:.. Nevertheless, sources do warrant *

for the conclusion that the main branches of the Ganges,

now called the Bhagirathi and the Padma, were in exis­

tence from at least about the dawn of the Christian era

or even from an earlier period.

Though most important, the Ganges and the Padma

are, however, not the only significant rivers of Bengal.

Rivers connected with as well as independent of them are

noteworthy in number and size.

It is only natural that in a plain as well as

deltaic Bengal rivers will not continue to flow in the

same course for ever. According to an observed geogra­

phical law, "Every stream whether large or small flowing

through such flat-level, tends to raise its bed or channel

by deposition of the silt and sand ... and by this gra­

dual deposition the channel bed of the streams is, raised

above the actual level of the adjoining flats. ... The

stream necessarily abandons its original course and seeks

a new channel in the lower ground adjoining, until after

:; 43 ::

successive changes it has gradually wandered over the

whole flat."

This may well have been Jthe natural phenomenon

responsible for the throwing up of the distributaries by

- 71
the Bhagirathi flowing through the deltaic Bengal.

It has been observed that silt and sand carried

down by the river had gradually raised its bed. When the

bed was raised above the surrounding level, the“accumula­

ted waters naturally sought for a new channel. The Bhai-

rava, one of the branches, might have originated in this

way. There are, however, indications that once it'

70. Proceedings ASB , 1870, P. 47.

The remarks of Somerville are interesting in this

connection : "The course of all rivers changes when they

pass from one geological formation to another, or by

dislocations of the strata. The sudden deviations in their

directions are generally ov/ing to the latter circums­

tances" (Physical Geography. P. 255).

71. K. Bagchi, The Ganges Delta, P. 52.

72. Ibid
:: 44 ::

served as one of the main outlets of the Gangetic waters.
At present, however, it is not a very significant branch
and is dissected into three parts by the Jelangi and the
Mathabhanga. It used to branch off from the Padma opposite
Rampur-Boalia and emptied itself into the Haringhata
74 .

Gradual silting up of river-beds and the consequent

rise of the adjoining levels led to the development of
other distributaries and re distributaries of the Ganges,
such as the MathabhSnga, the Kumar, Nabagahga» Chitra

Of these channels, the history of the KumSr is more

traceable than that of any other. Branching off from the
Mathabhanga, at present it joins the Garai and the com­
bined outfall being known in different parts as the Carai,
Madhumati, Silaidaha and the Baleswar discharges into the
sea through the Haringhata.

According to Fennell*s atlas the Kumar flowed

across the jftorth of the districts of Nadiya, Jessore and
Faridpur. A little below its junction with the Garai it
threw off a channel called Barasia which flowed southwards,

73. Q. Pi cat.
74. Ibid. P.54.
75. Ibid. P. 53*

• •
• • 45

while the parent:stream continued its course eastward

t 76
towards Padma. The Edilpur copper-plate of.Sricandra

mentions Kumara-tslaka-mandala and this political

t, • *"
suh-division is supposed to have*been geographically
associated with the Kumar.

x *

Existence of the aforesaid stream can perhaps

be traced back even to Ptolemy's time (2nd century A.D.)

if not earlier,He mentions the Kamberikhon as the third

mouth of the Ganges, and, it has been identified with
the Kumar-Baringhata outfall. It is thus possible that

in the second century A.D. it formed one of the main

channels of the Ganges.

Another important river of Central Bengal in the

mediaeval period was the Yamuna. It has now silted up... In

Rennell's map it appears to be a very narrow stream. In

the fifteenth century, however, in the days of Vipradasa,

the Yamuna was a mighty river. The Pavanaduta of Dhoyl
also mentions it as issuing out of the Bhagirathl.

76. El, XVII, P. 190.

77. EBR, Vol.I, P. 196.

78. Ptolemy / VII. I. 18.

79. HBR, Vol.I, P.ll.

80. Vipradasa, Manasa -Mane a la , P. 142.

81. Pavanadutam, V. 23.

: : 46 ::

* *>

It is apparent that the Yamuna began to deteriorate some­

time between the 16th and the 18th centuries, but the exact
. *

period is not known to us.

( C )

The Teesta, an important feature of the hydro­

graphy or North Bengal, now flows through inter alia the

districts of Jalpaiguri and Rangpur. In Rennell's map

the Teesta is shown as flowing through the Punarbhava,

Atreyi and Karatoya.’These streams.combined lower down

and taking the name Hoorsagar discharged into the Ganges,

at Juffarganj near the Goalundo,

The Hoorsagar is still living, being the combined

outfall of the Atreyi, Karatoya, Boral and the Yamunesvari.

But it falls into the Jamuna much above the confluence

at Goalundo.

The fact that the Teesta flowed through three

channels might be the reason for the ascription to her

of the name Trisrots (Abbreviated into Teesta) meaning

'three streams*1'* We may add here that in Rennell's map

a very narrow stream is shown as Tista-.c( creek) issuing

out of the Teesta a few miles below the town of Jalpaiguri.

82. D._R. Bhandarkar Volume, P. 343.


It flowed South-east through the district of Rangpur

and discharged into the Brahmaputra near Kaliganj. As
indicated above, it is through this course that the
Teesta flows at present. The '’destructive floods" of-1787
might have been responsible for this hydrographic change.

The Karatoya to-day is almost a dead river. Its

course is, however, well delineated in Rsnell’s Atlas.
According to it the Karatoya flowed first jSouth and then
South-east past Ghoraghat, Shibgunj, Mustanagar and Bogra

and met the Atreyl to the South of Shahjadpur. In the

maps of Jao de Barros and Kantelli de Vinela (1683) a river
flowing direct iriorth to south is called Caor. It flowed
through the kingdom of Reino de Komotali and another stream
flowing horth-west joined it immediately below that kingdom.
Reino de Komotah corresponded to Rangpur and Co&ch Behar.
So the river flowing through this state may be identified
with the Karatoya, whereas the second channel may represent
the upper course of the Brahmaputra. Such a suggestion
is strengthened by the fact, pointed out by Hunter, that

83. Hunter, op.cit. vol. VII, P, 165.

It is possible that the present Teesta also was an old
bed of the river. Vanc den Broiieke’s map represents the
Teesta as an affluent of the Brahmaputra and Rennel’s
Teesta C (Creek) represents the Teesta at present.
; • 48 •

the people living on the banks of the Karatoya believe

that it is identical with the Brahmaputra*

Travernier (1666) speaks.of a large river named

Chativor flowing from the jtforth. The name of the river

may suggest its identity with the Karatoya. In the map of

Van den Broucke (1660) the Karatoya is a much wider stream.

According to Mirza Nathan it flowed past Ghoraghat and in

the rainy season the river was "impassable”.

In the Rama carita of Sandhyakaranandin the

Karatoya is said to have formed the Eastern boundary of

Varendri. Hsiian Tsang in the seventh century A.D.

crossed over a mighty river while travelling from

u. 88
Pun-na- fa-tan-na (Pundravardhana) to Kia-mo-lo-po (KamavjPa).

84. Hunter, op. cit. P. 167.

86. Travenier's Travels in India* (BKI), P. 185.

86. Mirza Nathan, Baharistan-i-ghyabi. P. 53.

The river Karatoya is stated' to have' been very small

at Balia. It indicates that already by the beginning of

the seventeenth century the lower course of the Karatoya

corresponded to that indicated in Kennel's map. Balia has

been identified withBawleah, located in Rennel's Sheet

no.6, seven miles to the south-west of Shahjadpur.

87. Rama car ita , III, 10.

88. Watters, op. cit. Vol. II, P. 185.

In the Tang-Shu the river between these countries is
called Ka-lo-tu. Watters identified Ka-lo-tu with the
Brahmaputra. But the expression Ka-lo-tu is clearly a
Chinese corruption of the name Karatoya.
91 .
The Puranas and the tirthayatra section of the
Mahabharata glorify the sacredness of the Karatoya.

The Atrai at present branches off from the

Punarbhava and joins the J'amuna in the Rajshahi district.
According to Rennell's Atlas it flowed due South by the
side of Dinajpore, Puttyram and Jangipur* Flowing South­
east through the Chalanbil, it joined the Karatoya below
Shajadpur. The map of Van Den Broucke shows a small
stream issuing out of the Atrai and emptying itself into
the Padma.

Another mighty river of Worth Bengal was the

Punarbhava. In the last, century it united with the Maha-
nanda at Aiyarganj and the combined waters fell into the
Padmeu From Renneli's 'map it seems that once the Mahananda

89. Q-plait.TP. 187.

90. Ibid.
91. Padmapurana. XXI5 Msrkandeya Purana. LI, 21 - 25.
92. Mi* , HI, 85j Yogini tantra, I, II, 60i 1, 12,69.
:: 50

discharged itself into the Karatoya. The latter river thus

carried enormous amount of water, the outfall of,a number

of rivers of North Bengal.

It is a pity that with tiie exception of the

Karatoya the history of none of the above rivers of North

Bengal can be traced beyond the later mediaeval age. It

is also noteworthy that there are remarkable differences

between the present courses of these rivers and those

indicated in late mediaeval sources. The reason behind

the great changes'within a comparatively short span may

have been the devastating flood of 1787. It seems to have

changed the course of the Teesta. Since then, it appears,

this river failed to supply waters to the sister streams.

This naturally led to the rapid silting up of the beds of

the three rivers, viz. , the Karatoya, Atreyi and the Punar-
bhava. Gradually they were probably reduced to their

present straits.

93. Major F.C.Hirst (Report on the Nadiya Rivers. 1915:1


Appendix A), pointed out that in 1824. Lt. Fisher was

ordered to inespect the old course of the Brahmaputra

from the Jenai and report why the former was smothering;;

away. It was thus a difference of not less than forty years

between the change of the course of the Teesta on the one

hand and the deterioration of the Brahmaputra on the other.

Major Hirst came to conclusion that between 1787 and 1820
the Jenai slowly increased in importance and from 1850,
sixty-three years after the Teesta flood it had become the
main channel of the Brahmaputra.

( D )

The Brahmaputra, the mighty river of Worth and


South-east Bengal rises in the Manasa Sarohara in Tibet

and then running eastward 1 it takes a vicious turn into

Assam.Flowing through Assam it enters Bengal through th'e

borders of Rangpur and CoochBehar districts. The main

volume of its water then flows into the Jamuna district ?

from the river mentioned above. The joint stream flowing

through the borders of the districts of Bogra and Pabna

on the one hand, and those of Mymensingh and Dacca, on

the other, joins the Padma. The latter unites with the

Meghna between the districts of Faridpur and Noakhali and

then discharges into the Bay of .Bengal. Another channel

still known as the Brahmaputra comes out of the Jamuna

near the place of their meeting. It then flows through

the Mymensingh district and unite with the Meghna. The

joint stream debouches into the Bay of Bengal through the

opening between the districts of Woakhali and Buckergunj.

According to the map of Rennel, the Brahmaputra

skirting along the G-aro hills, ran South-eastward from

Dewanganj, it flowed past Sherpur, Bygonbani, through the

district of Mymensingh. Flowing east of the Madhupur jungle

and its Eastern part of the Dacca district it united with

the Meghna near Sunerampur in Tippera district.

The streams are, however, shown in this map as

"branching off from the Brahmaputra; one a little above

its confluence with the Meghna, the other a little below

the point where Re melt's Brahmaputra and the present

Jamuna bifurcate. The- former is the Lakkia and the lattey

is described as the old bed of the Brahmaputra both dis­

charging into the Dhaleswari. The holy bathing place of

Lungalbund is on this old bed of the Brahmaputra.

We can thus trace at-least three full or partial

courses of this river: (1) the present main course through

the Jamuna ; (2) the course through the Mymensingh district

and (3) the old bed of the Brahmaputra in Rennells’s Atlas.

There are indications that the present main

course of the Brahmaputra was the principal, if not only

channel in early mediaeval period.

, 94
Yasodhara (13th century), in his commentary on

the Kamasutra of Vatsyayana, places Vanga to the East of "

the Lauhitya (identified with the Brahmaputra). This has

generally been taken to indicate that Vanga atleast in

94. Karmasu/tra , P. 295.

95. Cf. Vanga-La uhityat Pur vena.

the thirteenth Century A.D., denoted the territory to the
east of the Jamuna, i. e. , the districts of Dacca, Mymen-
sing, etc. But a considerable portion of this tract lies
West to the course of the Brahmaputra -as delineated in
Rennel.'l’s map.

Hence if we attach any value to the information

of Yasojdhara, it is better to assume that his Lauhitya
was the Jamuna, the present main course of the Brahmaputra.

We admit that it is highly doubtful whether

Yasodhara had a clear conception of the geographical
denotation of. Vanga. The Madanpada (Kotalipada Parganah
of the Faridpur district) Copper-plate of Visvarupasena
(13th Century A.D.) records the grant of the village of
Pinjokasthi situated in Vikramapura-bhaga in Vanga within
Pundravardhana-bhukti. The provenatfi^f the inscription
may suggest that the Kotalipada region was a part of
Vanga in the thirteenth Century A.D.

It appears that the Brahmaputra in the early

mediaeval period flowed through the Jamuna channel.
Sometime later it flowed atleast partially through the
old bed referred to the Renneli's map. By Rennel.i’s time

54 t:

the river had began to flow through the Mymensing district.

Ultimately, however, its main volume of water began to

flow again through the Jamuna channel.

The Saktisangamatantra states that the entire
territory from the sea up to the Brahrflaputra denoted Vanga.

A Mandasore inscription of a much earlier period alludes

to the Lauhitya.

As has already been observed the change of the

course of the Teesta is supposed to have had been respon­

sible for the fluctuation of the course of the Brahmaputra

via the Jamuna at present. Major Hirzti’s report is that

in 1824 Fisher was ordered to inquire why the old Brahma­

putra was deteriorating. So there is a difference of about

40 years between the Teesta flood and the time when the

drying up of the old Brahmaputra was drawn into notice;

Hirst thinks that between 1787 and 1820 the Jamuna gradua­

lly increased in volume and importance and from 1850 it

began to serve as the main channel of the Brahmaputrs.

But in that case also the flood of 1787 and the fluctua­

tion of the course of the Teesta cannot be regarded as

being totally isolated from the shifting of the course of

the Brahmaputra.

98. Saktisangamatantra, Pf. Ill, Ch. VII, verse 17.

(Gos, no. civ, P.67).
99. Cf. Alauhityppakanthat tala-vana- gahanopatyakat in the
r" "'' l” ' "• * ' 1 1
Mandasore inscription of Xasodharman, ClI, III ,PP. 142-48.

All tile rivers of Western Bengal: excepting the

Sarasvati, originate from the Chojtanagpur plateau and hear

the characteristics of hill streams. Important rivers of West

Bengal are the Ajai, Damodar, Darakes-^ar, Kasai the Rupnarayan


The DSmodar at present enters Bengal at Barakar and

flows South-east past Raniganj, Andal and Burdwan. A little

helow the last place it suddenly takes a curve and flows due
South into the Hooghly a little above the Bhagirathi-Rupnarayan,

In Kennel's map (1764-77) the old course of the Damodara

is shown as flowing fcorth-east from about Burdwan and finally .

discharging into the Bhaglrathi a few miles above Hooghly,

while the actual course of the Damodara during Rennel's Survey

corresponds to the present channel of the river.

The map of Van <den Broucke (1660), however, shows the

Damodar as flowing eastward from Burdwan into the Bhaglrathi

near present Kalna. Ksemananda (1640) also refers to this course

’ 101
in his 'Manasar-Bhasan'.

The old course of the Damodar as shown in Rennell's

map is possibly represented by the Behula which flowing direct

east from the Damodar falls into the Bhagirathi a little above

100. Rennel, Bengal Atlas, Sheet no. 7.

101. Ksemananda, ed. by Bijan Bihari Bhattacharyya, P. 23.

:: 56 :: #

Hooghly. Similarly the Banks which flows a little 'Worth-east

from the DHmodar into the Bhagirathi near about Kalna5may
represent the course of the Damodar as shown in Van 3en
Broucke's map. This course is indeed called Banka DSmodar
by Ketakadasa Ksemananda.

In Rennell's map the course of the Rupnarayan is

shown as corresponding to its present bed. As has already
been observed in early ages the Rupnarayan united with the
Ganges a little above Tamralipta or Tamluk. At present,
however,.these two rivers join a little below Tamluk. This
hydrographic change was probably due to the shifting of the
Sarasvati Brahch of the Ganges.

The present Ajai fairly represent the course of

the river as shown in Kennel's atlas.

Of great antiquity is the river Kansai. At

present it flows in a semi-circular way past Midnapore
and then runs parallel to the Rupharayan up to the sea.
In Van den Broucke's map, however, the river on which
Midnapore is situated is shown as flowing into the

102. O.plcit.
55 57 :: *


Rupnarayan near Nannger. In Rennell's nap Narajigur is

situated on the river flowing South of the Kasai and
ultimately joining the latter. The combined fall of these
rivers is now known as the Haldi.

103 r _
Kalidasa refers to a river called Kapisa. In

the fourth canto of the Raghuvamsa it is stated that Raghu

crossed over this river to Utkala. The implication is that
Kapisa served as the eastern boundary of Utkala* This
river has been identified with the present Kansai. H.C.
Raychaudhuri suggests that the easternmost mouth of the
five months of Ptolemy's Ganges represents the Kansai. It
is difficult to say how far Dr. Raychaudhuri is correct.
We can only observe that the present Kansai has no conn­
ection with the Ganges; atleast it was not a distributary
of the latter.

It is quite clear from the above discussion that

the hydrography of Bengal has changed in different ages.

103. Raghuvamsa, IV, 37.