in the summer, late spring, or early fall, the most intense floods generally occur in winter, late fall, or early spring. These two agencies-storm and temperature are, as we have said, entirely outside the power of man to control. Science has discovered no means of modifying them in the slightest degree. Even the prediction of storms is confined to their progress and probable results after they are well developed. In the two great controlling factors in the production of floods man is thus entirely helpless, and his own work begins only after emergency has arisen.

II

Let us now consider the effect of man's occupancy of the earth — the process of replenishment - upon the destructiveness of floods. This opens up the chief controversial aspect of the problem - the alleged effect of man's operations in hastening the run-off of storm-water to the streams and in diminishing the natural capacity of the streams to carry it away. Deforestation, drainage, and encroachment upon flood-channels are the substance of the indictment. So wide is the scope of these matters that we can here enter into no argument concerning them, but must confine ourselves to a categorical summary of what we understand to be the conclusions of the engineering profession in its practical dealings with the problem of floodcontrol.

First, as to deforestation. It is held by forestry enthusiasts that the forest soil, by virtue of its cover, is more receptive of moisture than the open country. This has never been demonstrated and is very doubtful as a fact. Cultivation of the soil, ploughing, sowing, and so forth, greatly increase capacity for absorption. On the other hand, the undisturbed condition of the forest soil

through long periods, and the solid packing of the ground around the roots of large trees, tend toward impermeability of forest soils. Where the balance of these influences lies, it is impossible to say, but it is probably slight, one way or the other.

It is held that forest cover is the best protection against erosion. There is no protection superior to a well-knit sod, or a thin covering of close-standing crops or underbrush. There are situations where forest cover is probably best for this purpose, but there is no universal rule to that effect.

It is held that forests, by their cooler status, as compared with the open country, induce greater precipitation. As a factor in the flood-problem, even if this were so, it is insignificant and on the side of greater floods.

Beyond question, forests tend to intensify floods from melting snow. This is because they prevent drifting, and thus expose greater surfaces to melting influences.

The definite controlling fact in the whole matter, however, is this: that, even if there be a certain reservoir effect in forest cover, it invariably becomes exhausted in the long rains which lead up to great floods, and is altogether ineffectual when the crisis arrives.

The records of stream-flow where longest kept do not indicate an increase in the intensity or frequency of floods as a result of deforestation, or a contrary effect where reforestation has been long in progress.

Finally, whatever merit there may be in the conventional forest theory, it is impossible of general application because of the necessities of human existence. On the average, not more than one fourth of the land area can be given over to forests in a thickly settled country. That is less than the existing areas of virgin forest and re

It is necessary to lay emphasis on the foregoing facts in order to disabuse the public mind of an illusory theory which amounts almost to an obsession. So long as the public believes that it can protect itself from floods by planting trees, so long will the cause of effective flood-control suffer. How great is the need of enlightenment on this subject may be appreciated when so prominent a man as the late United States Senator Newland, of Nevada, who claimed to have special acquaintance with these subjects and was influential in legislation concerning them, was capable of an utterance like the following:

'Why, we have been destroying our forests, those great natural reservoirs of moisture into which the waters fall from the heavens, and where they are stored in the leaves and the loose soil, and drunk up by the thirsty roots of the trees and vegetation, and thence the surplus gradually makes its way to the creeks and the tributaries of our rivers. We have destroyed our forests, and the water which used to be absorbed by these forests is now hurried on into the creeks and the rivers.'

Secondly, as to drainage. It seems obvious enough that drainage workspavements, sewers, road-ditches, tile and open drainage on farms - must hasten storm-water to the streams. That may at once be accepted as a fact. But in this, as in forestry, there are powerful compensations which never occur to one on first thought. This, in particular, is the case with the drainage of lands naturally wet and marshy. In a state of nature these lands are thoroughly saturated in their normal condition. But drainage takes this water out, and creates enormous space for ground-storage where none existed before. So it happens that the

flow of water when once in the ditches, is ever operating to create ground-storage which shall delay the rapid filling of the ditches. We do not know where the balance lies, but our progressive friends, the French, lay especial stress on the restraining influence just pointed out.

Thirdly, as to channel encroachment. In this connection two kinds of channels are to be considered one, the normal channel between banks in which the ordinary flow of the stream is confined; the other, the overflow channel through the bottoms between the uplands on either side, which comes into use only in time of high water. These bottom lands are universally subject to encroachment and occupancy because they are the most valuable of all lands. This forces an extra burden upon normal channels in time of flood. As to these channels, it seems to be a fact (contrary to the accepted view) that man's work tends to increase rather than diminish their capacity. There are, of course, exceptions, some of them very pronounced; but the rule seems to be as just stated. A most striking example is furnished by the recent studies of the great Miami River flood-problem. The channel capacity through those of the chief towns on that stream where there have been radical modifications by human action, was found to be from two to eight times as great as in the country districts above and below, where man's interference has been slight. Such a condition cannot, of course, be merely an accident.

We may here observe that one of the most capricious and uncertain things in nature is the channel capacity of streams. It is subject to no uniform rule. On the same stream it may vary all the way from the magnitude of the greatest flood to perhaps not one per

cent of it; and so it results that the problem of flood-control, even on the same stream, may be entirely different on the different sections.

The conclusion of this whole question of the influence of man's occupancy of the earth upon the intensity and destructiveness of floods may be stated thus:

So far as deforestation, drainage, and so forth, are concerned, the compensating influences of cultivation are so many and important that we cannot tell where the balance lies. There is no reason to suppose that in great and prolonged floods it falls decisively on one side or the other. Whatever it may amount to, it is more apparent near the sources of streams than on the lower courses of great rivers, where it is subject to certain other influences to which we shall refer in discussing the subject of reservoirs.

The great and controlling effect of man's work in causing flood destruction arises from his occupancy of natural overflow channels (bottom lands), thus blockading Nature's chief highway for conducting her flood waters to the sea. It is here that man dwells in greatest numbers; it is here that wealth accumulates in greatest abundance. It is a situation in which man is in part responsible for his own misfortunes, and on him rests the burden of adequate provision for escape from them. What such provision should be, is our next inquiry.

III

This subject we may consider under three headings-warnings, prevention, protection.

Flood warnings. Although science. cannot predict the occurrence or intensity of a storm with the least degree of certainty, it can forecast something of its progress after it has begun, and it

can do a great deal to warn the public of its probable effects in run-off. The Flood Service, as this branch of public activity is called, thus becomes a matter of great importance, and in France and Germany it is developed to a high state of perfection. By its forecasts it can avert practically all loss of life and, to a large extent, loss of property. It is a highly useful department of the public service, and should be made as ef ficient as possible.

Flood-prevention. This feature of flood-control has to do with retarding the run-off so that it will not pass downstream as rapidly as it would naturally. It is altogether a matter of reservoir action-catching the water in great basins as it runs from the land, holding it until the storm is past, and letting it out gradually afterward. The method is so logical in theory that the enthusiast never stops to inquire what may be the limitations of its universal application; but, as a matter of fact, these are many and important. The subject itself, like that of forestry, is so extensive in its scope that argument pro and con is impossible in a paper of this length, and it is necessary therefore to stop with a detailed statement of the conclusions generally accepted by the engineering profession.

In their purpose and functioning. reservoirs are of two classes: storage reservoirs, in which the water is held for some later use, such as power, municipal supply, irrigation, and the like, flood-control being rather an incidental than a main purpose; and detention, or retarding, reservoirs, with outlets always open, so that the water detained during a storm is promptly released afterward, flood-control being in this case the sole consideration. In the storage reservoir the land occupied is permanently lost to other use. In the detention reservoir it remains available for agricultural use.

Inas

With the storage reservoir there is generally a conflict of purpose between flood-control and other uses. much as the quantity of rainfall in the wet season, when water is being collected, can never be foretold, it becomes important, for storage purposes, to fill the reservoirs as soon as possible, so as to be sure of a supply; while, for flood-control, it is important to reserve ample space until the season of storms is safely past. The conflict can best be harmonized by building the reservoir so large that it can safely store a maximum run-off; but this may greatly increase the cost and may often be impracticable for lack of site. In the majority of cases, both purposes cannot be completely satisfied.

The influence of reservoirs upon flood-control is, of course, greatest on those sections of the streams which lie immediately below the dams. The protection there afforded may be absolute. But the effect diminishes rapidly downstream as additional tributaries come in, and in large part disappears on the lower courses of great rivers like those of the Mississippi basin. This is due in part to the impracticability of building reservoirs on all the tributaries, but mainly to the impossibility of so manipulating outflow as to make it diminish floods on the main stream. Such floods are the result of combinations of floods from many tributaries. Reservoirs, particularly of the detention type, might release their stored waters at the wrong time, so far as the combination below is concerned. There can be no doubt that this would often be the case, and for that reason any system of reservoirs which could not hold back the flood water above them until the season of floods was safely past would be of doubtful value. The enormous development of the system necessary to make it at all effectual, and the difficulty of so manipulating

it as to reduce tributary combination, cause engineers to look with misgiving upon the scheme as a means of controlling the floods on the lower courses of great rivers like the Mississippi.

It seems clear that the reservoir principle of flood-control is bound to have wider and wider application; but it is doubtful whether this will take the form of detention basins for flood-control exclusively, so much as of storage reservoirs with flood-control as one consideration only. Flood-protection. This is distinguished from flood-prevention in that it does not seek to hold back the runoff, but rather to expedite its progress to the sea, and to make the channels carry it without overflow. The whole purpose is to increase channel capacity. This is accomplished by one of the following measures or by a combination of two or more of them:

(a) By cutting off bends, thus shortening the channel and increasing its slope, streams may be made to carry more water for the same dimensions. Cut-offs are a very important resource on small streams, and even on some large rivers, like those of the Hungarian plain. Official sentiment is strong against them in the case of the Mississippi, but for other reasons than those of flood-control.

(b) Enlargement of the channel itself may be effected by widening (which is least desirable because it costs more for right-of-way, requires much excavation, and restricts the area of bottom land available for occupancy); or by deepening, which involves expensive excavation and is likely to have to be repeated more than once; or, finally, by enlarging the channel upward through the building of dikes along the banks. This is the levee system - the cheapest, most effective, and most universally used of all methods of channel enlargement. It has been resorted to

(c) On some rivers it seems impossible ever to confine the greatest floods to the main channel. In such cases auxiliary flood channels, or bi-passes, have to be provided. The Sacramento is the most prominent example. It is considered desirable to avoid this method whenever it is possible to do so, and it has been definitely rejected on the Mississippi, although it has always had its advocates there.

(d) A bi-pass, strictly speaking, returns to the main stream. An outlet differs from it in leading to the sea by its own channel. All deltaic rivers have outlets near their mouths. From the point of view of navigation, they are looked upon with disfavor because they tend to dissipate the energy of the current and lead to the deterioration of the main channel. On the lower Mississippi one such natural outlet, the Atchafalaya, has been maintained; all the others have been closed. Surveys have been made for another in the vicinity of New Orleans, but official sentiment is generally against them for that stream.

IV

It is manifest, from this very cursory survey, that the flood-problem is one of great complexity, not to be worked out by any definite rule, but to be studied from every angle in order to determine the method or combination suited to the particular case. It is, perhaps, fortunate, in spite of the losses suffered, that there has been delay in solving the problem on most streams,

because technical knowledge on the subject has greatly increased in recent years, and the engineering profession is far better prepared than it was twenty, or even ten, years ago to deal with it effectively.

If we were to attempt to formulate any broad conclusion as to the application of these various measures of floodcontrol, it would be this: as a rule, reservoirs will find their greatest usefulness near the headwaters of streams; levees will be the main resource on the lower courses of great rivers; cut-offs will generally be confined to the smaller streams; and excavation will be resorted to only in special situations, where less expensive methods fail to accomplish the purpose.

There is the satisfying fact about the general problem of flood-control that it is capable of definite solution. Some of the railroads have so far provided against such disasters that they no longer fear them. Every thoroughgoing treatment places so much of the problem behind us. The total cost, moreover, may not prove so great as we are apt to think. In casually reviewing some of our more important problems, it is difficult to see how the aggregate for this country, in addition to that portion which will be defrayed on other accounts (through works for industrial use, and the like), can equal that of the Panama Canal. Spread over a quarter of a century, the burden will be light and the resulting benefits will virtually cancel it as we go along.

Perhaps the most perplexing question confronting the engineer who is called upon to formulate a scheme of flood-control in any particular case is to determine what degree of control to advise. He invariably becomes convinced that there is nothing in the conditions of the case which can furnish assurance that no greater flood will ever occur than those which he knows to