Rivers, as every body has seen, are always broadest at the mouth, and narrower towards their source. But what is less known, and probably more deserving curiosity, is, that they run in a more direct channel as they immediately leave their sources; and that their sinuosities and turnings become more numerous as they proceed. It is a certain sign among the savages of North America, that they are near the sea, when they find the rivers winding, and every now and then changing their direction. And this is even now become an indication to the Europeans themselves, in their journeys through those trackless forests. As those sinuosities, therefore, increase as the river approaches the sea, it is not to be wondered at, that they sometimes divide, and thus disembogue by different channels. The Danube disembogues into the Euxine by seven mouths; the Nile, by the same number; and the Wolga, by seventy.

There are some rivers which are said to lose themselves in chasms under the earth, and to flow for several miles in secret and undiscovered channels. On this circumstance is founded one of the most beautiful fables of antiquity, relative to the fountain of Arethusa, in Sicily. The same thing is affirmed of the Rhine, and even of the river Mole, in Surry, which from this circumstance derives its name. With respect to the two latter rivers, however, some doubts are entertained of the aserted fact.

On this subject there is a valuable article in the Memoirs lately published, by the abbé Guettard. "It is very surprising (he observes) if we reflect on it, that a river in its course, which is very often very extensive, should not meet with spongy soils to swallow up its waters, or gulphs in which they are lost; nevertheless, as there has been hitherto known but a small number of rivers whose waters thus disappear, this phenomenon has been accounted very extraordinary, both by the ancients and moderns. M. Guettard next describes what he has observed in several rivers of Normandy, which are lost and afterwards appear again; these are five in num❤ ber, viz. the Rille, the Ithom, the Aure, the river of Sap André, and the Drôme. The three first disappear gradually, and then come in sight again; the fourth loses itself entirely by degrees, but afterwards re appears; the fifth loses some of its water in its course, and ends by precipitating itself into a cavity, whence it is never seen to rise again.

What seems to occasion the loss of the Rille, the Ithon, and the

Aure, is the nature of the soil through which they pass. M. Guet tard has observed that it in general porous, and composed of a thick sand, the grains of which are not well compacted together; it sinks suddenly down by its own weight in some places, and there forms great holes; and when the water overflows the meadows, it fre quently makes many cavities in several parts of them. If we there fore suppose inequalities in the channels of these rivers, and that there are certain places in which the water stagnates longer than in others, it must there dilute the ground, if we may use that expres sion; and having carried away the parts which united the grains of sand together, those grains will become afterwards no other than a kind of sieve, through which the waters will filtrate themselves, provided nevertheless that they find a passage under-ground through which they may run. This conjecture appears to be so well founded, that each of these three rivers loses itself nearly in the same man. ner, that is, through cavities which the people of the country call betoirs, and which swallow up more or less according to their largeness. M. Guettard, who has carefully examined them, remarks, that these betoirs are holes in the form of a tunnel, whose diameter and aperture is at least two feet, and sometimes exceeds eleven; and whose depth varies in like manner from one and two feet, to five, six, and even twenty. The Rille during the summer season loses almost all its water in the space of two short leagues; the Ithon does very near the same. But M. Guettard observes something curious concerning this river, that formerly it was not lost, but kept its course without any interruption, as appears by the history of the country; very likely the mud, which had been collected together in several parts of its channel, might have occasioned the waters remaining in others, and have caused many betoirs. This is the more likely, as the mud having been collected together in the bed of the river Aure, it appears that, in consequence, the cavities were greatly increased, which makes it lose itself much sooner than formerly. Besides, possibly an earthquake happening in the country might have caused several subterraneous canals through which the water of the Ithon has forced its way. In effect, it appears, that a soil's being porous is not sufficient to cause the loss of a river; for if it was, then to do so it would occasion many fens round about, nor would it renew its course after having disappeared a certain time; it must

besides find ways under-ground through which it may take its course. M. Guettard seems also much inclined to believe, that there are, in these parts, subterraneous cavities through which the waters may flow; and in consequence of this he reports a number of facts, all tending to prove the truth of it, or at least to prove that there must be hollow quarries serving for strainers to these waters. Upon which occasion he goes into a discussion of this question; Are there any subterraneous rivers, and is the prepossession of some persons in favour of this particular well founded? He makes it appear by several instances which he quotes, and by many reasons which he alleges, that there are at least very great presumptions in favour of this opinion. We are too apt not to look beyond the exterior of things: we feel resistance upon the surface of the earth; when we go deep, we often find it it compact. It is therefore hard for us to imagine that it can contain subterraneous cavities. sufficient to form channels for hidden rivers, or for any considerable body of water; in a word, that it can contain vast caverns; and yet every thing seems to indicate the contrary. A fact that is observed in the betoirs of the rivers concerning which we have spoken, and particularly of the Rille, proves in some measure that there are considerable lakes of waters in the mountains which limit its course; this fact is, that in winter the greatest part of their betoirs become springs, which supply anew the river's channel with as much water as they had absorbed from it during the summer. Now from whence can that water come, unless from the reservoirs or lakes that are inclosed in mountains, which being lower than the river in summer, absorb its waters, and being higher in winter by the rain they receive, send it back again in their turn?

M. Guettard strengthens this conjecture by several instances that render it very probable: he remarks at the same time, that this alternate effect of the betoirs swallowing up the water and restoring it again, causes perhaps an invincible obstacle to the restraining of the water within the channel of the river. It has indeed been several times attempted to stop those cavities; but the water returns with such violence in winter, that it generally carries away the materials with which they were stopped.

The river of Sap André is lost in part, as we have before said, in the same manner as the Ithon and the Rille; but there is something

more remarkable in it than in those rivers; to wit, that at the extremity of its course, where there is no perceptible cavity, it is ingulphed, but without any fall; the water passes between the pebbles, and it is impossible to force a stick into that place any further than into the betoirs of which we have spoken. What makes this river take that subterraneous direction, is an impediment which its stream meets with in that place; it is there stopped by a rising ground six or seven feet high, whose bottom it has very likely undermined, to gain a free passage, not having been able to make its way over it. At some distance it appears again; but in winter, as there is a greater quantity of water, it passes over that eminence, and keeps an uninterrupted course.

Lastly, the Drôme, after having lost some of its water in its course, vanishes entirely near the pit of Soucy; in that place it meets with a sort of subterraneous cavity near 25 feet wide, and more than 15 deep, where the river is in a manner stopped, and into which it enters, though without any perceptible motion, and never appears again.

M. Guettard finishes this memoir with some observations upon the Ierre. This river is lost in the same manner as the Rille; and though it is very near Paris, this singularity is unknown to almost every body; was it not for the account of M. l'Abbé le Boeuf, M. Guettard would have been also ignorant of it. And as he thinks the chief object of a naturalist's observation ought to be the public good, he examines the means which might be employed to restrain the water of the Terre. The same object has made him add a description of the manner how the Rhône is lost, or rather how its course is disturbed; for it is now very certain that it does not lose itself, but that its channel is extremely confined, in the place where it was pretended that it lost itself, by two mountains, between whose feet it runs. M. Guettard makes it appear that it might not be impossible to widen that place, and give a sufficient channel to the river, which would render it navigable, and be of vast utility to all the country."

Pantologia, Art. Rivers. Phil. Trans. Year 1690.
Mem. de l'Acad. des Sciences.

SECTION II.

Principal Rivers in the different Quarters of the World.

THE general course of the largest rivers we are acquainted with is from a thousand to two thousand miles; and we have them of this length in every quarter of the world; yet in no instance do we find them much exceeding two thousand miles long. We shall commence our rapid tour with those of

ASIA.

The rivers that here attract our attention are the Indus and the Ganges.

The INDUS is by the natives called Sinde or Sindet, and in the Sanscrit language Seendho. It is also called Nilab, or the Blue River. The source, both of this and of the Ganges, are to this hour unknown: Major Rennell, and various other geographers, have of fered opinions upon the subject, but at present they are opinions and nothing more. It is generally supposed to originate in the mountains of Mus Tag, which, as laid down by Strahlenburg, run from west to east, forming a chain to the south of Little Bucharia. Its comparative course may be about a thousand British miles, when it forms a Delta in the province of Sindi, entering by numerous mouths into the Indian sea.

The tributary streams of the Indus chiefly join it in the northern half of its course, where they form the Panjab, or county of Five Rivers. From the west run into Indus the Kamet, with its auxiliary streams, and the Comul; from the east the Bahut or Hydaspes; the Chunab or Acesinas; the Kauvee or Hydraotes; and the Setlege or Hesudrus, being on the east of the Indus. The whole of this part of Hindustan is even at present but little known to the moderns; and it is uncertain whether the Caggan, a considerable and distant river to the east, joins the Indus or falls into the gulph of Cuteh.

The GANGES is a still nobler stream, both in magnitude and length; for it is swelled by tributary streams of still greater number and power, and its comparative length can be scarcely estimated at