This translation contains quotations from the works of M. Belgrand and other French engineers, who had made the hydrology of the basin of the Seine a special study. Among other things M. Belgrand says:

This country comprises all or part of 21 Departments, as follows: Aisne, Ardennes, Aube, Cote d'Or, Eure, Eure-et-Loire, Loiret, Marne, Haute-Marne, Nievre, Nord, Oise, Pas-de-Calais, Seine, Seine-Inferieure, Seine-et-Marne, Somme, Vosges, and Yonne, and comprises an area of about 107,000 square kilometers, nearly equal to the fifth part of the area of country comprising the 86 Departments.

The most irregular streams, those most subject to rapid rises, are found especially in the Departments of the Yonne, Nievre, and Cote d'Or, and to a less extent in those of the Aube, Haute-Marne, and Aisne. This region is very woody, more so, perhaps, than the rest of France. The most remarkable Departments in regard to the regularity of the water courses which rise within them are the Eure, Eure-et-Loire, Nord, Oise, Pas-de-Calais, Seine-Inferieure, Somme, the chalky parts of the Aube and of the Marne, and those portions of the Seine-et-Oise and Loiret in which the limestones of La Beauce abound. The majority of the streams in these countries are subject to slight rises of short duration, their stage of water varying but little. This group of Departments is perhaps one of the most sparsely wooded in France, because the Eure, Eure-et-Loire, Nord, Oise, Pas-de-Calais, Seine-Inferieure, and the Somme have only about one-tenth of their surface wooded, and the plateaus of La Beauce and the chalky plains of Champagne are, if we except some recent plantations of pine, completely bare of trees.

In order to test the question as to the effect of forests in regulating the flow of water M. Belgrand had daily measurements made from November, 1850, to May, 1853, of the discharges of the Cousin and of the Grenetierre, which is one of its affluents. Both of these basins are of granite formation, impermeable and otherwise alike, but the first is only about one-third wooded, while the second is entirely covered with trees. Notwithstanding this great difference as regards the extent of forests in each, the results have been the same in both, as is shown by the following account:

"The regimen of each is identically the same, although their valleys are unequally wooded. Their waters rise and fall at the same rate, whether in rainy weather or in dry, in winter or in summer; their low winter regimen is more abundant than that of summer.

"A heavy rain in winter produces in both a sudden flood of greater or less height, but of very short duration, followed by a long stage of tolerably high water; the sudden and high freshets take place in each at the same time.'

The different details concerning the flow of water are, then, exactly the same in the two basins, and yet one is entirely covered with forests, while in the other twothirds is bare of trees. M. Belgrand has made a number of more detailed observations yet, which show, further, that it is not upon forests but upon cultivated ground that the greatest regularity in flowage is observed.

In Vallès's paper he quotes from a report on the basin of the Eure made by M. St. Clair, engineer in chief, showing the beneficial effects of cleared and cultivated lands in diminishing by absorption the amount of surface water, as follows:

All the valleys, even those of least extent, are cut up by ravines which were often formerly the beds of torrents. Within the last twelve years the condition has changed; they are now almost always dry. The cause of this great change, the progress of agriculture, is generally recognized in the country. The soil has been cultivated more and rendered more permeable; the farmers, reaping more advantages from the culture of the ground, and fully aware of the utility of improving it, have, by means of ditches, hedges, and endikements properly located, controlled the flow of water everywhere, preventing erosion, causing fertilizing deposits of sediment, and relieving the surface of the ground from the asperities which interfered with cultivation. The waters, retarded thus in their flow, have settled in great part through the ground and disappeared before reaching the ravines.

These agricultural improvements, in a country where land susceptible of cultivation, amounts to sixty-one one-hundredths of the area of the country, have, then, reduced the surface flowage and increased the absorption.

• Annales des Ponts et Chaussees, annee 1852, premier semestre, p. 102.
26320-10-2

They have rendered less frequent and formidable the freshets, which, in the basin of the Eure, are to be attributed to an excess of surface water rather than to any supply from springs, which latter is almost always invariable.

The effect upon springs of cutting down forests can now be easily eliminated. We will divide them into two classes, viz, superficial and subterranean.

The first issue from the points of the surface very near to the strata in which the waters which produce them collect. The second, on the contrary, are found very far below these strata, and the water, in order to find an outlet, traverses frequently long distances underground.

The first, generally small, pertain indifferently to absorbent or nonabeorbent ground; but the second, occasionally very powerful, pertain essentially and almost exclusively to permeable soils.

Now, it is indisputable that the continual humidity of the soil of forests is favorable to the first and ought to maintain in their feeble flow considerable regularity. It is, then, very likely that the clearing away of forests and exposing the earth to alternations of drought and moisture would alter the regimen of these springs; that these would be more abundant in time of rain; that they would decrease in summer and possibly be dry for several months in the year. This explains the disappearance of certain springs after the cutting down of forests.

As regards the second group, which are plentifully supplied by infiltration through the permeable coils, it is different.

From the different manner in which, as regards absorption, wooded and cultivated soils act, we see that in the first this faculty is in great part destroyed, while in the second it is increased. To remove standing timber from permeable soils is to restore to them the facility of transmitting the waters which_the_forest vegetation, whether by the spreading of its roots, by the fall of leaves, or by the compactness of the soil, had taken from them, and it results, consequently, in a more abundant supply of water to the subterranean springs.

Thus, it is worthy of remark that the most abundant of all of them, especially in seasons of low water, are located beneath the vast ledges of limestone which are almost entirely denuded; for instance, those of Cahors and Louysse, in the department du Lot, and the famous fountain of Vaucluse, of which mention has already been made.

M. Belgrand further says:

Now these basins, so remarkably alike, we have; they are those of the Seine in the seventeenth, eighteenth, and nineteenth centuries. Everything in these is alike excepting the extent of the forests, which has steadily decreased, so that if we were in possession of adequate information of some exact measurements of the greatest inundations during the time specified, we could easily test the correctness of our theories.

In fact, observations of this nature have been made; they go back to 1615, about the time when French industry began to develop, and when, consequently, the felling of timber to a great extent commenced.

This places at our disposal an interval of five half centuries.

In a memoir published in 1814 by the engineer Egault, these observations were compiled, discussed, and arranged with reference to the heights of the most marked inundations. We add to the results collated by them those which have been obtained since his time and give them in the following table:

The deductions from this table are striking. The continued decrease of the floods for each half century is remarkable. The waters attained a mean height of 27.53 feet in the first half of the seventeenth century; they only attained a mean of 21.22 feet in the present. According to this, we have experienced an amelioration of nearly 6.56 feet, and yet the trees have been steadily and unceasingly cut down, and the forests transformed into cultivated farms.

What would we gain, then, to-day, I ask, in rewooding our field? It would be but an unfortunate attempt to restore the old order of things, when the floods of the Seine rose to 29.53 feet above the low stage.

In connection with the conclusions reached in this report, as well as with regard to those reached by the foresters and others who differ from my views, I would emphasize the fact that none of us have flood data extending over any great period of time, but in Europe we fortunately have some long-period observations. The preceding pages show the result of observations made by competent engineers during two and one-half centuries in the basin of the Seine, and show that there has been a gradual and constant decrease in the height of floods with the diminution of forests.

In Germany another long-period record is presented. Mr. Ernest Lauder, chief of the hydrographic bureau of the Austrian Government, recently made an exhaustive investigation of the records of the Danube, the great river of central Europe. He prepared an exhaustive report on the destructive floods in the Danube that occurred in 1897 and 1899, and in this report traces the history of the floods of the Danube for eight hundred years, taking into account 125 different floods. His conclusions are that progressive deforestation of the country has had no effect in increasing the frequency of floods or in augmenting their height. Among other things he showed that the flood of 1899, which was a summer flood, was severest where it came from the heavily wooded districts.

Much has been written about the barren condition of the valley of the Jordan, in the Holy Land, and it is pointed out that great cities and teeming populations once covered the regions now barren; but this does not prove that if there has been a decrease in the rainfall it is due to deforestation, for everywhere in this region are evidences of extensive irrigation that was practised at the time this region. was thickly populated. The date palm, the vine, and the fig tree will grow there as luxuriantly to-day as in the old Biblical days, if artificial irrigation is used, as was formerly done. It is not believed that the cutting of the cedars of Lebanon has had anything to do with the dryness of the adjacent regions.

At the tenth International Congress of Irrigation, held at Milan in 1905, papers were presented by representatives from France, Germany, Italy, Austria, and Russia, in which the writers heartily favored the protection of the forests and their cultivation. But these writers were unanimous in the opinion that forests exercise little influence upon either the high water or the low water of rivers. In this connection I will quote from Col. H. M. Chittenden, M. Am. Soc. C. E., volume 34, page 944, Proceedings of the Society of Civil Engineers, as follows:

The constantly reiterated statement that floods are increasing in frequency and intensity, as compared with former times, has nothing to support it. There are, it is true, periods when floods are more frequent than at others, and hasty conclusions are always drawn at such times; but, taking the records year after year for considerable periods, no change worth considering is discoverable. The explanation of these periods of high water, like the one now prevailing, must, of course, be sought in pre

cipitation. That is where floods come from, and it is very strange that those who are looking so eagerly for a cause of these floods jump at an indirect cause and leave the direct one entirely untouched. In the records of precipitation, wherever they exist, will be found a full and complete explanation of every one of the floods that have seemed unusually frequent and severe in recent years.

THE SOURCE OF FLOOD WATERS IN THE UNITED STATES.

Before one can get a comprehensive idea of the magnitude of the problem involved in the creation of the floods of the United States, it will be necessary for him to first study chart A, which gives a typical illustration of the cyclonic storms that frequently form on the Rocky Mountain Plateau, either on its northern, central, or southern portions. Under the influence of gravity air flows from regions where the pressure is great toward the regions where it is less. In the case illustrated by this chart the atmosphere, as indicated by the direction in which the arrows point, is flowing from the region marked "high," which is central over the Carolinas, toward the region where the pressure is low, which is central over Montana, and the vaporous atmosphere that rises from the Gulf of Mexico and the adjacent ocean is carried far into the interior of the continent. Conditions similar to these occur many times each month, and as a result the eastern and central portions of the United States are bathed in a succession of rains which, as shown by chart B, gradually thin out and largely disappear on the eastward edge of the Rocky Mountain Plateau, because the currents of air from the Gulf of Mexico do not reach farther inland.

STATEMENT BY MR. BAILEY WILLIS.

In the May issue, 1909, of the magazine entitled "Conservation," page 262, Mr. Bailey Willis makes the statement:

The moisture which falls upon North America in the form of rain and snow comes chiefly from the Pacific Ocean. A smaller proportion, rising from the Gulf of Mexico and the West Indian seas, falls upon the eastern United States.

It is true that chart B, giving the normal annual precipitation indicates that the Pacific Ocean furnishes precipitation that is heavy along the immediate coast, but that it is the principal source, as Mr Willis says, of the moisture that falls upon the North American Continent, is not borne out by the facts exhibited by the precipitation chart herein produced and by the inflowing currents of air that are shown on chart A.

The range of mountains on the Pacific coast intercepts the inflow of the vaporous atmosphere, which is comparatively shallow, and precipitates its aqueous vapor on the windward side of the range, and mainly on the north half of the windward side, because storms seldom enter from the southern half. To be sure, some of the scant precipitation that falls on the plateau does drift over the tops of these mountains, but the amount is small. Certain it is that the Pacific Ocean has little influence on the precipitation of the eastern half of the United States, which fact is well understood by meteorologists; and I believe that most of them will join me in the belief that the only way that man could materially affect the rainfall of the eastern half of the United States would be to erect a mountain barrier