[352] In some cases where the bed of rapid Alpine streams is composed of very hard rock—as is the case in many of the valleys once filled by ancient glaciers—and especially where they are fed by glaciers not overhung by crumbling cliffs, the channel may remain almost unchanged for centuries. This is observable in many of the tributaries of the Dora Baltea, which drains the valley of the Aosta. Several of these small rivers are spanned by more or less perfect Roman bridges—one of which, that over the Lys at Pont St. Martin, is still in good repair and in constant use. An examination of the rocks on which the abutments of this and some other similar structures are founded, and of the channels of the rivers they cross, shows that the beds of the streams cannot have been much elevated or depressed since the bridges were built. In other cases, as at the outlet of the Val Tournanche at Chatillon, where a single rib of a Roman bridge still remains, there is nothing to forbid the supposition that the deep excavation of the channel may have been partly effected at a much later period. See App., No. 47.

[353] Mémoire sur les Inondations des Rivières de l'Ardèche, p. 16. "The terrific roar, the thunder of the raging torrents proceeds principally from the stones which are rolled along in the bed of the stream. This movement is attended with such powerful attrition that, in the Southern Alps, the atmosphere of valleys where the limestone contains bitumen, has, at the time of floods, the marked bituminous smell produced by rubbing pieces of such limestone together."—Wessely, Die Oesterreichischien Alpenländer, i, p. 113. See Appendix, No. 48.

[354] Frisi, Del modo di regolare i Fiumi e i Torrenti, pp. 4-19.

[355] Surell, Étude sur les Torrents, pp. 31-36.

[356] Champion, Les Inondations en France, iii, p. 156, note.

[357] Notwithstanding this favorable circumstance, the damage done by the inundation of 1840 in the valley of the Rhone was estimated at seventy-two millions of francs.—Champion, Les Inondations en France, iv, p. 124.

Several smaller floods of the Rhone, experienced at a somewhat earlier season of the year in 1846, occasioned a loss of forty-five millions of francs. "What if," says Dumont, "instead of happening in October, that is between harvest and seedtime, they had occurred before the crops were secured? The damage would have been counted by hundreds of millions."—Des Travaux Publics, p. 99, note.

[358] Troy, Étude sur le Reboisement des Montagnes, §§ 6, 7, 21.

[359] For accounts of damage from the bursting of reservoirs, see Vallée, Mémoire sur les Reservoirs d'Alimentation des Canaux, Annales des Ponts et Chaussées, 1833, 1er sémestre, p. 261.

[360] Some geographical writers apply the term bifurcation exclusively to this intercommunication of rivers; others, with more etymological propriety, use it to express the division of great rivers into branches at the head of their deltas. A technical term is wanting to designate the phenomenon mentioned in the text.

[361] Mardigny, Mémoire sur les Inondations de l'Ardèche, p. 13.

[362] In the case of rivers flowing through wide alluvial plains and much inclined to shift their beds, like the Po, the embankments often leave a very wide space between them. The dikes of the Po are sometimes three or four miles apart.—Baumgarten, after Lombardini, Annales des Ponts et Chaussées, 1847, 1er sémestre, p. 149.

[363] It appears from the investigations of Lombardini that the rate of elevation of the bed of the Po has been much exaggerated by earlier writers, and in some parts of its course the change is so slow that its level may be regarded as nearly constant.—Baumgarten, volume before cited, pp. 175, et seqq. See Appendix, No. 49.

If the western coast of the Adriatic is undergoing a secular depression, as many circumstances concur to prove, the sinking of the plain near the coast may both tend to prevent the deposit of sediment in the river bed by increasing the velocity of its current, and compensate the elevation really produced by deposits, so that no sensible elevation would result, though much gravel and slime might be let fall.

[364] To secure the city of Sacramento in California from the inundations to which it is subject, a dike or levée was built upon the bank of the river and raised to an elevation above that of the highest known floods, and it was connected, below the town, with grounds lying considerably above the river. On one occasion a breach in the dike occurred above the town at a very high stage of the flood. The water poured in behind it, and overflowed the lower part of the city, which remained submerged for some time after the river had retired to its ordinary level, because the dike, which had been built to keep the water out, now kept it in.

According to Arthur Young, on the lower Po, where the surface of the river has been elevated much above the level of the adjacent fields by diking, the peasants in his time frequently endeavored to secure their grounds against threatened devastation through the bursting of the dikes, by crossing the river when the danger became imminent and opening a cut in the opposite bank, thus saving their own property by flooding their neighbors'. He adds, that at high water the navigation of the river was absolutely interdicted, except to mail and passenger boats, and that the guards fired upon all others; the object of the prohibition being to prevent the peasants from resorting to this measure of self-defence.—Travels in Italy and Spain, Nov. 7, 1789.

In a flood of the Po in 1839, a breach of the embankment took place at Bonizzo. The water poured through and inundated 116,000 acres, or 181 square miles, of the plain, to the depth of from twenty to twenty-three feet in its lower parts.—Baumgarten, after Lombardini, volume before cited, p. 152.

[365] Moyens de forcer les Torrents de rendre une partie du sol qu'ils ravagent, et d'empêcher les grandes Inondations.

[366] The effect of trees and other detached obstructions in checking the flow of water is particularly noticed by Palissy in his essay on Waters and Fountains, p. 173, edition of 1844. "There be," says he, "in divers parts of France, and specially at Nantes, wooden bridges, where, to break the force of the waters and of the floating ice, which might endamage the piers of the said bridges, they have driven upright timbers into the bed of the rivers above the said piers, without the which they should abide but little. And in like wise, the trees which be planted along the mountains do much deaden the violence of the waters that flow from them."

[367] I do not mean to say that all rivers excavate their own valleys, for I have no doubt that in the majority of cases such depressions of the surface originate in higher geological causes, and hence the valley makes the river, not the river the valley. But even if we suppose a basin of the hardest rock to be elevated at once, completely formed, from the submarine abyss where it was fashioned, the first shower of rain that falls upon it after it rises to the air, while its waters will follow the lowest lines of the surface, will cut those lines deeper, and so on with every successive rain. The disintegrated rock from the upper part of the basin forms the lower by alluvial deposit, which is constantly transported farther and farther until the resistance of gravitation and cohesion balances the mechanical force of the running water. Thus plains, more or less steeply inclined, are formed, in which the river is constantly changing its bed, according to the perpetually varying force and direction of its currents, modified as they are by ever-fluctuating conditions. Thus the Po is said to have long inclined to move its channel southward in consequence of the superior mechanical force of its northern affluents. A diversion of these tributaries from their present beds, so that they should enter the main stream at other points and in different directions, might modify the whole course of that great river. But the mechanical force of the tributary is not the only element of its influence on the course of the principal stream. The deposits it lodges in the bed of the latter, acting as simple obstructions or causes of diversion, are not less important agents of change.

[368] The distance to which a new obstruction to the flow of a river, whether by a dam or by a deposit in its channel, will retard its current, or, in popular phrase, "set back the water," is a problem of more difficult practical solution than almost any other in hydraulics. The elements—such as straightness or crookedness of channel, character of bottom and banks, volume and previous velocity of current, mass of water far above the obstruction, extraordinary drought or humidity of seasons, relative extent to which the river may be affected by the precipitation in its own basin, and by supplies received through subterranean channels from sources so distant as to be exposed to very different meteorological influences, effects of clearing and other improvements always going on in new countries—are all extremely difficult, and some of them impossible, to be known and measured. In the American States, very numerous watermills have been erected within a few years, and there is scarcely a stream in the settled portion of the country which has not several milldams upon it. When a dam is raised—a process which the gradual diminution of the summer currents renders frequently necessary—or when a new dam is built, it often happens that the meadows above are flowed, or that the retardation of the stream extends back to the dam next above. This leads to frequent lawsuits. From the great uncertainty of the facts, the testimony is more conflicting in these than in any other class of cases, and the obstinacy with which "water causes" are disputed has become proverbial.

The subterranean courses of the waters form a subject very difficult of investigation, and it is only recently that its vast importance has been recognized. The interesting observations of Schmidt on the caves of the Karst and their rivers throw much light on the underground hydrography of limestone districts, and serve to explain how, in the low peninsula of Florida, rivers, which must have their sources in mountains a hundred or more miles distant, can pour out of the earth in currents large enough to admit of steamboat navigation to their very basins of eruption. Artesian wells are revealing to us the existence of subterranean lakes and rivers sometimes superposed one above another in successive sheets; but the still more important subject of the absorption of water by earth and its transmission by infiltration is yet wrapped in great obscurity.

[369] The sediment of the Po has filled up some lagoons and swamps in its delta, and converted them into comparatively dry land; but, on the other hand, the retardation of the current from the lengthening of its course, and the diminution of its velocity by the deposits at its mouth, have forced its waters at some higher points to spread in spite of embankments, and thus fertile fields have been turned into unhealthy and unproductive marshes.—See Botter, Sulla condizione dei Terreni Maremmani nel Ferrarese. Annali di Agricoltura, etc., Fasc. v, 1863.

[370] Deep borings have not detected any essential difference in the quantity or quality of the deposits of the Nile for forty or fifty, or, as some compute, for a hundred centuries. From what vast store of rich earth does this river derive the three or four inches of fertilizing material which it spreads over the soil of Egypt every hundred years? Not from the White Nile, for that river drops nearly all its suspended matter in the broad expansions and slow current of its channel south of the tenth degree of north latitude. Nor does it appear that much sediment is contributed by the Bahr-el-Azrek, which flows through forests for a great part of its course. I have been informed by an old European resident of Egypt who is very familiar with the Upper Nile, that almost the whole of the earth with which its waters are charged is brought down by the Takazzé.

[371] It is very probably true that, as Lombardini supposes, the plain of Lombardy was anciently covered with forests and morasses (Baumgarten, l. c. p. 156); but, had the Po remained unconfined, its deposits would have raised its banks as fast as its bed, and there is no obvious reason why this plain should be more marshy than other alluvial flats traversed by great rivers. Its lower course would possibly have become more marshy than at present, but the banks of its middle and upper course would have been in a better condition for agricultural use than they now are.

[372] From daily measurements during a period of fourteen years—1827 to 1840—the mean delivery of the Po at Ponte Lagoscuro, below the entrance of its last tributary, is found to be 1,720 cubic mètres, or 60,745 cubic feet, per second. Its smallest delivery is 186 cubic mètres, or 6,569 cubic feet, its greatest 5,156 cubic mètres, or 182,094 cubic feet.—Baumgarten, following Lombardini, volume before cited, p. 159.

The average delivery of the Nile being 101,000 cubic feet per second, it follows that the Po contributes to the Adriatic six tenths as much water as the Nile to the Mediterranean—a result which will surprise most readers.

[373] We are quite safe in supposing that the valley of the Nile has been occupied by man at least 5,000 years. The dates of Egyptian chronology are uncertain, but I believe no inquirer estimates the age of the great pyramids at less than forty centuries, and the construction of such works implies an already ancient civilization.

[374] There are many dikes in Egypt, but they are employed in but a very few cases to exclude the waters of the inundation. Their office is to retain the water received at high Nile into the inclosures formed by them until it shall have deposited its sediment or been drawn out for irrigation; and they serve also as causeways for interior communication during the floods. The Egyptian dikes, therefore, instead of forcing the river, like those of the Po, to transport its sediment to the sea, help to retain the slime, which, if the flow of the current over the land were not obstructed, might be carried back into the channel, and at last to the Mediterranean.

[375] The Mediterranean front of the Delta may be estimated at one hundred and fifty miles in length. Two cubic miles of earth would more than fill up the lagoons on the coast, and the remaining ten, even allowing the mean depth of the water to be twenty fathoms, which is beyond the truth, would have been sufficient to extend the coast line about three miles farther seaward, and thus, including the land gained by the filling up of the lagoons, to add more than five hundred square miles to the area of Egypt. Nor is this all; for the retardation of the current, by lengthening the course and consequently diminishing the inclination of the channel, would have increased the deposit of suspended matter, and proportionally augmented the total effect of the embankment.

[376] For the convenience of navigation, and to lessen the danger of inundation by giving greater directness, and, of course, rapidity to the current, bends in rivers are sometimes cut off and winding channels made straight. This process has the same general effects as diking, and therefore cannot be employed without many of the same results.

This practice has often been resorted to on the Mississippi with advantage to navigation, but it is quite another question whether that advantage has not been too dearly purchased by the injury to the banks at lower points. If we suppose a river to have a navigable course of 1,600 miles as measured by its natural channel, with a descent of 800 feet, we shall have a fall of six inches to the mile. If the length of channel be reduced to 1,200 miles by cutting off bends, the fall is increased to eight inches per mile. The augmentation of velocity consequent upon this increase of inclination is not computable without taking into account other elements, such as depth and volume of water, diminution of direct resistance, and the like, but in almost any supposable case, it would be sufficient to produce great effects on the height of floods, the deposit of sediment in the channel, on the shores, and at the outlet, the erosion of banks and other points of much geographical importance.

The Po, in those parts of its course where the embankments leave a wide space between, often cuts off bends in its channel and straightens its course. These short cuts are called salti, or leaps, and sometimes reduce the distance between their termini by several miles. In 1777, the salto of Cottaro shortened a distance of 7,000 mètres by 5,000, or, in other words, reduced the length of the channel more than three miles; and in 1807 and 1810 the two salti of Mezzanone effected a reduction of distance to the amount of between seven and eight miles.—Baumgarten, l. c. p. 38.

[377] The fact, that the mixing of salt and fresh water in coast marshes and lagoons is deleterious to the sanitary condition of the vicinity, seems almost universally admitted, though the precise reason why a mixture of both should be more injurious than either alone, is not altogether clear. It has been suggested that the admission of salt water to the lagoons and rivers kills many fresh water plants and animals, while the fresh water is equally fatal to many marine organisms, and that the decomposition of the remains originates poisonous miasmata. Other theories however have been proposed. The whole subject is fully and ably discussed by Dr. Salvagnoli Marchetti in the appendix to his valuable Rapporto sul Bonificamento delle Maremme Toscane. See also the Memorie Economico-Statistiche sulle Maremme Toscane, of the same author.

[378] This curious fact is thus stated in the preface to Fossombroni (Memorie sopra la Val di Chiana, edition of 1835, p. xiii), from which also I borrow most of the data hereafter given with respect to that valley: "It is perhaps not universally known, that the swallows, which come from the north [south] to spend the summer in our climate, do not frequent marshy districts with a malarious atmosphere. A proof of the restoration of salubrity in the Val di Chiana is furnished by these aerial visitors, which had never before been seen in those low grounds, but which have appeared within a few years at Forano and other points similarly situated."

Is the air of swamps destructive to the swallows, or is their absence in such localities merely due to the want of human habitations, near which this half-domestic bird loves to breed, perhaps because the house fly and other insects which follow man are found only in the vicinity of his dwellings?

In almost all European countries, the swallow is protected, by popular opinion or superstition, from the persecution to which almost all other birds are subject. It is possible that this respect for the swallow is founded upon ancient observation of the fact just stated on the authority of Fossombroni. Ignorance mistakes the effect for the cause, and the absence of this bird may have been supposed to be the occasion, not the consequence, of the unhealthiness of particular localities. This opinion once adopted, the swallow would become a sacred bird, and in process of time fables and legends would be invented to give additional sanction to the prejudices which protected it. The Romans considered the swallow as consecrated to the Penates, or household gods, and according to Peretti (Le Serate del Villaggio, p. 168) the Lombard peasantry think it a sin to kill them, because they are le gallinelle del Signore, the chickens of the Lord.

The following little Tuscan rispetto from Gradi (Racconti Popolari, p. 33) well expresses the feeling of the peasantry toward this bird:

O rondinella che passi lo mare
Torna 'ndietro, vo' dirti du' parole;
Dammi 'na penna delle tue bell' ale,
Vo' scrivere 'na lettera al mi' amore;
E quando l' avrò scritta 'n carta bella,
Ti renderò la penna, o rondinella;
E quando l' avrò scritta 'n carta bianca,
Ti renderò la penna che ti manca;
E quando l' avrò scritta in carta d' oro,
Ti renderò la penna al tuo bel volo.

O swallow, that fliest beyond the sea,
Turn back! I would fain have a word with thee.
A feather oh grant, from thy wing so bright!
For I to my sweetheart a letter would write;
And when it is written on paper fine
I'll give thee, O swallow, that feather of thine;
—On paper so white, and I'll give thee back,
O pretty swallow, the pen thou dost lack;
—On paper of gold, and then I'll restore
To thy beautiful pinion the feather once more.

Popular traditions and superstitions are so closely connected with localities, that, though an emigrant people may carry them to a foreign land, they seldom survive a second generation. The swallow, however, is still protected in New England by prejudices of transatlantic origin; and I remember hearing, in my childhood, that if the swallows were killed, the cows would give bloody milk.

[379] Morozzi, Dello stato antico e moderno del fiume Arno, ii, p. 42.

[380] Morozzi, Dello stato, etc., dell' Arno, ii, pp. 39, 40.

[381] Torricelli thus expressed himself on this point: "If we content ourselves with what nature has made practicable to human industry, we shall endeavor to control, as far as possible, the outlets of these streams, which, by raising the bed of the valley with their deposits, will realize the fable of the Tagus and the Pactolus, and truly roll golden sands for him that is wise enough to avail himself of them."—Fossombroni, Memorie sopra la Val di Chiana, p. 219.

[382] Arrian observes that at the junction of the Hydaspes and the Acesines, both of which are described as wide streams, "one very narrow river is formed of two confluents, and its current is very swift."—Arrian, Alex. Anab., vi, 4.

[383] This difficulty has been remedied as to one important river of the Maremma, the Pecora, by clearings recently executed along its upper course. "The condition of this marsh and of its affluents are now, November, 1859, much changed, and it is advisable to prosecute its improvement by deposits. In consequence of the extensive felling of the woods upon the plains, hills, and mountains of the territory of Massa and Scarlino, within the last ten years, the Pecora and other affluents of the marsh receive, during the rains, water abundantly charged with slime, so that the deposits within the first division of the marsh are already considerable, and we may now hope to see the whole marsh and pond filled up in a much shorter time than we had a right to expect before 1850. This circumstance totally changes the terms of the question, because the filling of the marsh and pond, which then seemed almost impossible on account of the small amount of sediment deposited by the Pecora, has now become practicable."—Salvagnoli, Rapporto sul Bonificamento delle Maremme Toscane, pp. li, lii.

The annual amount of sediment brought down by the rivers of the Maremma is computed at more than 12,000,000 cubic yards, or enough to raise an area of four square miles one yard. Between 1830 and 1859 more than three times that quantity was deposited in the marsh and shoal water lake of Castiglione alone.—Salvagnoli, Raccolta di Documenti, pp. 74, 75.

[384] The tide rises ten inches on the coast of Tuscany. See Memoir by Fantoni, in the appendix to Salvagnoli, Rapporto, p. 189.

On the tides of the Mediterranean, see Böttger, Das Mittelmeer, p. 190. Not having Admiral Smyth's Mediterranean—on which Böttger's work is founded—at hand, I do not know how far credit is due to the former author for the matter contained in the chapter referred to.

[385] In Catholic countries, the discipline of the church requires a meagre diet at certain seasons, and as fish is not flesh, there is a great demand for that article of food at those periods. For the convenience of monasteries and their patrons, and as a source of pecuniary emolument to ecclesiastical establishments and sometimes to lay proprietors, great numbers of artificial fish ponds were created during the Middle Ages. They were generally shallow pools formed by damming up the outlet of marshes, and they were among the most fruitful sources of endemic disease, and of the peculiar malignity of the epidemics which so often ravaged Europe in those centuries. These ponds, in religious hands, were too sacred to be infringed upon for sanitary purposes, and when belonging to powerful lay lords they were almost as inviolable. The rights of fishery were a standing obstacle to every proposal of hydraulic improvement, and to this day large and fertile districts in Southern Europe remain sickly and almost unimproved and uninhabited, because the draining of the ponds upon them would reduce the income of proprietors who derive large profits by supplying the faithful, in Lent, with fish, and with various species of waterfowl which, though very fat, are, ecclesiastically speaking, meagre.

[386] Macchiavelli advised the Government of Tuscany "to provide that men should restore the wholesomeness of the soil by cultivation, and purify the air by fires."—Salvagnoli, Memorie, p. 111.

[387] Giorgini, Sur les causes de l'Insalubrité de l'air dans le voisinage des marais, etc., lue à l'Académie des Sciences à Paris, le 12 Juillet, 1825. Reprinted in Salvagnoli, Rapporto, etc., appendice, p. 5, et seqq.

[388] See the careful estimates of Roset, Moyens de forcer les Torrents, etc., pp. 42, 44.

[389] Rivers which transport sand, gravel, pebbles, heavy mineral matter in short, tend to raise their own beds; those charged only with fine, light earth, to cut them deeper. The prairie rivers of the West have deep channels, because the mineral matter they carry down is not heavy enough to resist the impulse of even a moderate current, and those tributaries of the Po which deposit their sediment in the lakes—the Ticino, the Adda, the Oglio, and the Mincio—flow, in deep cuts, for the same reason.—Baumgarten, l. c., p. 132.

[390] "The stream carries this mud, &c., at first farther to the east, and only lets it fall where the force of the current becomes weakened. This explains the continual advance of the land seaward along the Syrian coast, in consequence of which Tyre and Sidon no longer lie on the shore, but some distance inland. That the Nile contributes to this deposit may easily be seen, even by the unscientific observer, from the stained and turbid character of the water for many miles from its mouths. A somewhat alarming phenomenon was observed in this neighborhood in 1801, on board the English frigate Romulus, Captain Culverhouse, on a voyage from Acre to Abukir. Dr. E. D. Clarke, who was a passenger on board this ship, thus describes it:

"'26th July.—To-day, Sunday, we accompanied the captain to the wardroom to dine, as usual, with his officers. While we were at table, we heard the sailors who were throwing the lead suddenly cry out: "Three and a half!" The captain sprang up, was on deck in an instant, and, almost at the same moment, the ship slackened her way, and veered about. Every sailor on board supposed she would ground at once. Meanwhile, however, as the ship came round, the whole surface of the water was seen to be covered with thick, black mud, which extended so far that it appeared like an island. At the same time, actual land was nowhere to be seen—not even from the masthead—nor was any notice of such a shoal to be found on any chart on board. The fact is, as we learned afterward, that a stratum of mud, stretching from the mouths of the Nile for many miles out into the open sea, forms a movable deposit along the Egyptian coast. If this deposit is driven forward by powerful currents, it sometimes rises to the surface, and disturbs the mariner by the sudden appearance of shoals where the charts lead him to expect a considerable depth of water. But these strata of mud are, in reality, not in the least dangerous. As soon as a ship strikes them they break up at once, and a frigate may hold her course in perfect safety where an inexperienced pilot, misled by his soundings, would every moment expect to be stranded.'"—Böttger, Das Mittelmeer, pp. 188, 189.

[391] The caves of Carniola receive considerable rivers from the surface of the earth, which cannot, in all cases, be identified with streams flowing out of them at other points, and like phenomena are not uncommon in other limestone countries.

The cases are certainly not numerous where marine currents are known to pour continuously into cavities beneath the surface of the earth, but there is at least one well-authenticated instance of this sort—that of the mill streams at Argostoli in the island of Cephalonia. It had been long observed that the sea water flowed into several rifts and cavities in the limestone rocks of the coast, but the phenomenon has excited little attention until very recently. In 1833, three of the entrances were closed, and a regular channel, sixteen feet long and three feet wide, with a fall of three feet, was cut into the mouth of a larger cavity. The sea water flowed into this canal, and could be followed eighteen or twenty feet beyond its inner terminus, when it disappeared in holes and clefts in the rock.

In 1858, the canal had been enlarged to the width of five feet and a half, and a depth of a foot. The water pours rapidly through the canal into an irregular depression and forms a pool, the surface of which is three or four feet below the adjacent soil, and about two and a half or three feet below the level of the sea. From this pool it escapes through several holes and clefts in the rock, and has not yet been found to emerge elsewhere.

There is a tide at Argostoli of about six inches in still weather, but it is considerably higher with a south wind. I do not find it stated whether water flows through the canal into the cavity at low tide, but it distinctly appears that there is no refluent current, as of course there could not be from a basin so much below the sea. Mousson found the delivery through the canal to be at the rate of 24.88 cubic feet to the second; at what stage of the tide does not appear. Other mills of the same sort have been erected, and there appear to be several points on the coast where the sea flows into the land.

Various hypotheses have been suggested to explain this phenomenon, some of which assume that the water descends to a great depth beneath the crust of the earth, but the supposition of a difference of level in the surface of the sea on the opposite sides of the island, which seems confirmed by other circumstances, is the most obvious method of explaining these singular facts. If we suppose the level of the water on one side of the island to be raised by the action of currents three or four feet higher than on the other, the existence of cavities and channels in the rock would easily account for a subterranean current beneath the island, and the apertures of escape might be so deep or so small as to elude observation. See Aus der Natur, vol. 19, pp. 129, et seqq. See Appendix, No. 53.

[392] "The affluents received by the Seine below Rouen are so inconsiderable, that the augmentation of the volume of that river must be ascribed principally to springs rising in its bed. This is a point of which engineers now take notice, and M. Belgrand, the able officer charged with the improvement of the navigation of the Seine between Paris and Rouen, has devoted much attention to it."—Babinet, Études et Lectures, iii, p. 185.

On page 232 of the volume just quoted, the same author observes: "In the lower part of its course, from the falls of the Oise, the Seine receives so few important affluents, that evaporation alone would suffice to exhaust all the water which passes under the bridges of Paris."

This supposes a much greater amount of evaporation than has been usually computed, but I believe it is well settled that the Seine conveys to the sea much more water than is discharged into it by all its superficial branches.

[393] Girard and Duchatelet maintain that the subterranean waters of Paris are absolutely stagnant. See their report on drainage by artesian wells, Annales des Ponts et Chaussées, 1833, 2me sémestre, pp. 313, et seqq.

This opinion, if locally true, cannot be generally so, for it is inconsistent with the well-known fact that the very first eruption of water from a boring often brings up leaves and other objects which must have been carried into the underground reservoirs by currents.

[394] Physikalische Geographie, p. 286. It does not appear whether this inference is Mariotte's or Wittwer's. I suppose it is a conclusion of the latter.

[395] Physical Geography of the Sea. Tenth edition. London, 1861, § 274.

[396] Paramelle, Quellenkunde, mit einem Vorwort von B. Cotta, 1856.

[397] Études et Lectures, vi, p. 118.

[398] "The area of soil dried by draining is constantly increasing, and the water received by the surface from atmospheric precipitation is thereby partly conducted into new channels, and, in general, carried off more rapidly than before. Will not this fact exert an influence on the condition of many springs, whose basin of supply thus undergoes a partial or complete transformation? I am convinced that it will, and it is important to collect data for solving the question." Bernhard Cotta, Preface to Paramelle, Quellenkunde (German translation), pp. vii, viii. See Appendix, No. 54.

[399] See the interesting observations of Kriegk on this subject, Schriften zur allgemeinen Erdkunde, cap. iii, § 6, and especially the passages in Ritter's Erdkunde, vol. i, there referred to.

Laurent, (Mémoires sur le Sahara Oriental, pp. 8, 9), in speaking of a river at El-Faid, "which, like all those of the desert, is, most of the time, without water," observes, that many wells are dug in the bed of the river in the dry season, and that the subterranean current thus reached appears to extend itself laterally, at about the same level, at least a kilomètre from the river, as water is found by digging to the depth of twelve or fifteen mètres at a village situated at that distance from the bank.

The most remarkable case of infiltration known to me by personal observation is the occurrence of fresh water in the beach sand on the eastern side of the Gulf of Akaba, the eastern arm of the Red Sea. If you dig a cavity in the beach near the sea level, it soon fills with water so fresh as not to be undrinkable, though the sea water two or three yards from it contains even more than the average quantity of salt. It cannot be maintained that this is sea water freshened by filtration through a few feet or inches of sand, for salt water cannot be deprived of its salt by that process. It can only come from the highlands of Arabia, and it would seem that there must exist some large reservoir in the interior to furnish a supply which, in spite of evaporation, holds out for months after the last rains of winter, and perhaps even through the year. I observed the fact in the month of June.

The precipitation in the mountains that border the Red Sea is not known by pluviometric measurement, but the mass of debris brought down the ravines by the torrents proves that their volume must be large. The proportion of surface covered by sand and absorbent earth, in Arabia Petræa and the neighboring countries, is small, and the mountains drain themselves rapidly into the wadies or ravines where the torrents are formed; but the beds of earth and disintegrated rock at the bottom of the valleys are of so loose and porous texture, that a great quantity of water is absorbed in saturating them before a visible current is formed on their surface. In a heavy thunder storm, accompanied by a deluging rain, which I witnessed at Mount Sinai in the month of May, a large stream of water poured, in an almost continuous cascade, down the steep ravine north of the convent, by which travellers sometimes descend from the plateau between the two peaks, but after reaching the foot of the mountain, it flowed but a few yards before it was swallowed up in the sands.

[400] It is conceivable that in large and shallow subterranean basins the superincumbent earth may rest upon the water and be partly supported by it. In such case the weight of the earth would be an additional, if not the sole, cause of the ascent of the water through the tubes of artesian wells. The elasticity of gases in the cavities may also aid in forcing up water.

A French engineer, M. Mullot, invented a simple method of bringing to the surface water from any one of several successive accumulations at different depths, or of raising it, unmixed, from two or more of them at once. It consists in employing concentric tubes, one within the other, leaving a space for the rise of water between them, and reaching each to the sheet from which it is intended to draw.