been, in its turn, the principal mouth; and the size of the delta itself has probably undergone, in the course of ages, material changes from the action of the waters, and the deposites left by the periodical floods.* It is thought that the Ganges is but little swelled by the melting of the snows, the waters derived from this source being not more than sufficient to balance the waste by evaporation; but it evidently owes its periodical rise in part to the rains which fall in the mountains. The sum total of its rise is 32 feet, out of which it rises 15 feet by the latter end of June; • "During eleven years of my residence in Bengal, the outlet or head of the Jellinghy river was gradually removed three quarters of a mile further down; and by two surveys of an adjacent bank of the Ganges, taken about the distance of nine years from each other, it appeared that the breadth of an English mile and a half had been taken away.... The windings of the Ganges in the plains, are doubtless owing to the looseness of the soil: the proof of it is, that they are perpetually changing....There are not wanting instances of a total change of course in some of the Bengal rivers. The Mootyjyl lake is one of the windings of a former channel of the Cossimbazar river. The Cosa river (equal to the Rhine) once ran by Purneah, and joined the Ganges opposite Rajemal: its junction is now 45 miles further up. Gour, the ancient capital of Bengal, stood on the old bank of the Ganges, although its ruins are four or five miles from the present bank. Appearances favour very strongly the opinion, that the Ganges had its former bed in the tract now occupied by the lakes and morasses between Nattore and Jaffiergungee; striking out of its present course at Bauleah, and passing by Pootyah. With an equal degree of probability, favoured by tradition, we may trace its supposed course by Dacca, to a junction with the Burrampooter (or Megna) near Fringybazar, where the accumulation of two such mighty streams probably scooped out the present amazing bed of the Megna."-RENNELL, pp. 343-6. These remarks may throw some light on the perplexed hydrography of the Punjaub. The quantity of land that has been destroyed by the Ganges in the course of a few years, between Colgong and Sooty, is estimated at 25,000 acres, or forty square miles. But the opposite shore has gained, and the new island of Sundeep contains above ten square miles. PART I. and the rainy season does not begin in most of the flat countries till about that time. In the mountains, the rains begin early in April; and by the end of that month, when the rain water has reached Bengal, the rivers begin very slowly to rise; at the rate, for the first fortnight, of an inch per day. This gradually augments to two or three inches, as the rain begins to reach the plains; and when it has become general, the mean increase is five inches per day. By the atter end of July, all the lower parts of Bengal contiguous to the Ganges and Brahmapootra, are overflowed, the inundation spreading more than 100 miles in breadth. Throughout this extent, nothing is seen but villages and trees, with here and there the top of an artificial mound, the site of some abandoned village. Owing to the great quantity of rain that falls in Bengal, the lands are, in general, overflowed to a considerable height, long before the bed of the river is filled, the ground adjacent to its banks, to the extent of some miles, being higher than the rest of the country.* There are particular tracts guarded from inundation by dikes, which are kept up at an enormous expense, but are not always effectual: these are cal *This is doubtless owing, as Buffon suggested, to the deposite left by the waters, which must be less in proportion to the distance from the margin of the river. The same circumstance has been remarked in the valley of the Nile. (See Mod. Trav., Egypt, vol. i. p. 42.) "Even when the inundation becomes general," Mr. Rennell says, "the river still shews itself, as well by the grass and reeds on its banks, as by its rapid and muddy stream; for the water of the inundation acquires a blackish hue by having been so long stagnant among grass and other vegetables. Nor does it ever lose this tint, which is a proof of the predominancy of the rain water over that of the river, as the slow rate of the motion of the inundation (which does not exceed half a mile per hour) is of the re markable flatness of the country."-RENNELL, pp. 349, 50. culated to extend in length, collectively, upwards of 1000 miles. By the end of July, the rate of increase has reached its maximum. In the first eight or ten days of August, it declines to four inches increase per day. During the second week in that month, the inundation in Bengal is nearly at a stand; and after the 15th, although great quantities of rain continue to fall throughout September, the waters begin to run off at the rate of from three to four inches, which gradually lessens, during November, to one inch; and from November to the end of April, the mean daily decrease is only half an inch. During the swoln state of the river, the tide, except very near the sea, totally loses its effect; but a strong wind blowing up the river for any continuance, will raise the waters two feet above their ordinary level, and the consequences are sometimes very disastrous.* The Ganges rises in a more considerable degree in the lower part of its course, than the northern rivers which communicate with it; the Burrampooter only excepted. Thus, during the dry season, the waters of the Teesta (which for about 150 miles has a course nearly parallel) flow into the Ganges by two distinct channels, while, by a third, it discharges part of its waters into the Burrampooter. But during the inundation, the Ganges runs into the Teesta, whose only outlet is into the Megna. From about the place where the tide * In ordinary seasons, the growth of the rice keeps pace with the rise of the waters, but it is destroyed by a too sudden rise. The harvest is often reaped in boats, as in some parts of South America.See MOD. TRAV., Brazil, vol. ii. p. 127, note. In 1763, in consequence of a strong gale conspiring with a high spring tide, at a season when the flood was within a foot and a half of its highest pitch, the waters rose six feet at Luckipoor, and the inhabitants of a considerable district were, with their houses and cattle, totally swept away,-RENNELL, p. 351, commences, to the sea, the height of the inundation gradually diminishes, till, at the point of confluence, it disappears.* Similar circumstances take place in the Burrampooter, the Meinam, and other rivers;† of which, Major Rennell suggests, the obvious explanation may be found in the known laws of fluids. The quantity of water discharged into the ocean by the Ganges, is computed by Mr. Rennell to be greater than that of any other river in the world; the mean quantity throughout the year being nearly 180,000 cubic feet in a second. The quantity of sand and soil held in suspension by its waters, is so great, that, in the year 1794, one of the mouths of the Bhagiruttee, at Sadigunge, full five miles in length, was, in the course of a week, filled up very nearly to a level with the contiguous country, although it must have contained about 900,000,000 solid feet. In the neighbourhood of Colgong, where the depth of the river is, in many places, upwards of 70 feet, new islands have risen to more than 20 feet above the level of the stream. ‡ The proper name of the Ganges, in the language of Hindostan, is said to be Pudda, which has been given to it as flowing ex pede Vistnou, from Vishnou's foot.§ But, in the Hindoo mythology, Gunga is a goddess, the daughter of Mount Himavut; and she is the object of worship with all castes. An annual festival is held in commemoration of her descent to the earth; offerings are then made to the river, and clay images of Gunga are set up in the temples, which are after * At Luckipoor, there is a difference of about six feet in the height of the river at different seasons; at Dacca, 14 feet; at Custee (240 miles from the sea by the river), 31 feet. + See Mod. Trav., Birmah, p. 7. Rennell, pp. 335-55. Hamilton's Hindostan, vol. i. p. 11. § Or, as some of the pooranus declare, ex sudore pedis. wards thrown into the Ganges. The waters are held to have an expiatory efficacy; and in so great reverence is this river held by the Hindoos, that in the British courts of justice throughout Bengal, witnesses are made to swear upon water from the Ganges, in the same manner as Christians are sworn upon the Gospels, and Mohammedans upon the Koran.* It is only that part of the river which lies in a line from Gangoutri to Sagor Island, that is deemed particularly sacred, and that is called the Gunga or Bhagiruttee. The great branch that runs eastward to join the Burrampooter, is not esteemed equally sacred; and it is to this that the name Pudda, Padma, or Padmawati, is generally restricted. The most sacred places are the prayags or junctions: of these, that of the Ganges and the Jumna, at Allahabad, termed simply Prayag, is deemed the principal. Hurdwar, where the river first escapes from the mountains, and Sagor Island, at the mouth of the Hooghly, are also sacred places. From Hurdwar to the sea, the Ganges forms a navigable stream 1350 miles in length, receiving, in its course, eleven rivers, some of which are equal to the Rhine, and none smaller than the Thames,+ be Ward's Hindoos, vol. i. p. 273. Many respectable Hindoos object, however, to this ordeal. †The principal of these are: 1. The Jumna, which, after a course of 780 miles, joins it on its right bank at Allahabad. 2. The Soane or Sona, which rises in Gundwana, and falls into the Ganges a little above Patna. 3. It next receives, on its left, the Ramgunga from Kemaoon. 4. The Goomty (Gomati, winding), rising in the same hills, crosses the province of Oude from N.W. to S.E., passing Luknow, and falls into the Ganges below Benares. 5. The Goggrah (Chahghara), one of its longest tributaries, separates Kemaoon from the Goorkha territory; at Swarga-dwara, joins the Sareyu or Sarjeu; and flowing through Oude, joins the Ganges in Bahar, under the name of Devar iver. 6. The Gunduk (called, in its higher parts, Salgrami), rises in the Himalaya, and after a course |