of the river ice plays but a small and insignificant part; some bank cutting is, of course, done by the floating ice pack, and some earth is brought down with the ice and dropped as it melts, but the amount of such cutting and filling is insignificant when compared with that due to other causes. In one respect only is the effect of ice at all important, and that is in the destruction wrought by it to the dikes and other regulation works, and in the interference with the floating plant engaged in building these works, but even in this respect the damage done by ice is far less than that caused by drift. This term "drift" is used to include every variety of floating timber from full-grown trees with unbroken roots and branches down to the smallest piece of wood. Some of this drift is brought down by tributary streams; some, such as saw logs, etc., is lost from rafts, but the larger part is dropped into the river by the caving of its banks. Except where cleared for cultivation, the entire bottom is heavily timbered, and when the river cuts into a timbered bank the timber, of course, falls into the river. Some of this floats away at once, but not all of it. Some of the trees slip in roots foremost, the tops falling over and leaning against the bank, while the roots. usually take with them a mass of earth that serves for a while to hold them down.

In time, of course, the earth around the roots is washed away, and with a rise in the river the tree floats and is carried downstream. This bank caving is going on at all stages, and consequently timber is getting into the river at all stages, but the drift that floats off at low water, or during a falling stage, is usually caught by a sand bar before it has traveled far, and consequently during low water and the falling stages the river is nearly or quite free from drift. During a rise the rate of bank caving is accelerated, and the timber floats off more freely, the drift that has been caught on low bars is taken up again and the whole continues its way downstream. Should the river fall again, much of the drift in motion is caught on sand bars or in chutes, to be floated off again at the next rise. The amount of drift in motion is therefore greatest during sudden rises, and at times there has been so much of it running as to cause the almost complete cessation of navigation for several days.

The total amount of drift in the river is very large, and it is said that in former times the planters along the lower part of the river depended entirely for their year's supply of fuel upon driftwood picked up as it floated by.

The effect of drift upon the regimen of the river is slight, but when we come to the consideration of regulation works it becomes of much greater importance. The floating drift is caught by a dike or similar structure and quickly accumulates. The drift coming against an already accumulated mass of drift is sucked by the current beneath it, and the drift raft thickens from below. In Plum Point Reach drift rafts have been formed 600 feet long, with a thickness of as much as 18 feet, and in some cases the drift raft has been massed up solidly to the bottom of the river. In cases of large accumulation of drift the trash caught in it acts to make the raft largely impervious, and the flow of water being checked, the energy of the current is converted into pressure, and if the dike be across a chute behind an island the velocity head is increased by that due to the fall in the main river between the head and foot of the chute. When it is considered that the drift raft covers all or nearly all of the upstream side of the dike, and transfers directly to the dike all this hydraulic pressure, it can be well understood that the force against the dike must at times be exceedingly great. In mattress work also much damage has been done by masses of drift getting under the mattresses during construction, and when much drift is running special means have to be taken to prevent trouble from it. As, however, the subject of the damage done by drift is to be afterwards referred to, in describing the works themselves, no further reference to it is necessary here.

MATTER CARRIED IN SUSPENSION

The water of this river is what is commonly called "muddy," being of a yellow color and containing in suspension more or less clay and fine sand. Many measurements have been made of the amount thus carried in suspension, and it has been found to vary greatly. It is generally greatest nearest the banks and bottom of the river, and least at the surface near the middle of the stream. It is usually greatest during high water and least in low water, and seems to vary from about 1 to 400 to 1 to 6,000 of the water by weight, and in general this proportion is greater the swifter the current, but this is not always true.

This matter is held in suspension by the motion of the water and has constantly a tendency to settle, and if the water comes to rest it will in time be deposited. This will be seen from the following fact: If a glassful of the water be caught, it will appear very muddy, but if it be not disturbed the settling begins at once and

in an hour or so the greater part of the suspended matter will have been deposited, though there will still be in suspension a quantity of exceedingly fine clay that will settle much more slowly. The depressions or lakes in the bottom land outside of the levees are, of course, in floods filled with muddy river water, but after the subsidence of the flood the water in these lakes soon clears up. At the lower ends of some high-water chutes, ponds or pools are left after high water. During the low-water season these, too, contain clear water, even though open at their lower ends. All these cases show that if the water becomes still the matter in suspension is deposited, and that it is kept in suspension by the motion of the water.

One other easily observable fact connected with sediment is important. On calm days, in long, regular bends of the river, where the banks are fairly regular and the water deep, and the irregu larities of the bottom not too great, and the motion of the water therefore fairly steady and regular, sedimentation begins right in the river itself and in the middle of the stream where the current is swiftest. The surface layer of the water begins to lose its load of suspended matter and to clear up, and the appearance of the surface is much the same as in a puddle which, after being stirred up, has been allowed to rest and has partially settled. With care a person in a skiff, floating with the current, can dip up some of this comparatively clear water, which is found to contain much less than the usual quantity of suspended matter. This shows an important fact, that it is not the main current of the river, or the motion of the water relative to the bank that keeps this matter in suspension, but the irregularities in this motion.

All the matter is carried in suspension contrary to the law of gravity, and with every opportunity it settles down, trying to get lower and lower and to deposit itself on the bottom. Some of it does, in fact, so deposit itself, and all of it would were it not for the eddies and whirls which keep it in suspension. By the caving of banks and cutting of the bottom new matter is constantly being taken into suspension near these banks and the bottom, and the irregularities of motion are constantly carrying it to the surface and to the middle of the river, and distributing it throughout the mass of water, and it is probable that if the river could flow through a channel smooth and regular on sides and bottom the irregularity of motion would cease and the matter held in suspension would shortly be either deposited or collected near the bottom

and sides and the water away from the bank and near the surface would be comparatively clear.

The fact just brought out, that the carrying of sediment depends upon the irregularities of motion of the river rather than upon the current itself, is important, for though these irregularities in motion will usually increase or decrease with an increase or decrease in the velocity of the current, they may not, and hence the irregularity of motion, and consequently the amount of material in suspension, will not vary in proportion to the velocity at different times and places, and there will therefore be no fixed relation between the velocity of the stream and amount of sediment it carries. (This last fact was early found by observation to be true, and from it certain erroneous ideas concerning river hydraulics were at one time drawn.)

In general, however, if at any place we reduce the velocity of the current, we reduce all kinds of motion in it, and consequently its power to transport matter in suspension, and a deposit will be made; and if we, on the other hand, increase the velocity of a current flowing close to the bank or bed of the river, we will increase the whirls and irregularities in it, and therefore its ability to carry matter, and a scour will be made.

In this connection it may be interesting to mention one peculiar feature to be seen on this river. At Cairo there is a distinct difference in color between the water coming from the Upper Mississippi and that from the Ohio, the former being relatively black and the latter yellow. As these two streams meet, these different colored waters begin to mix, but that process is slow, and under favorable circumstances the line of demarcation between the dark current on the right and the yellow on the left is distinctly noticeable 20 miles or more below Cairo.

MATERIAL MOVED ALONG THE BOTTOM

This consists usually of coarse sand and gravel too heavy to be taken into suspension. This motion is due to the excess of velocity pressure on the upstream side of the particle, and if this be sufficient the particles are torn from their places and rolled or pushed along the current, and in general the greater the velocity of the current at and near the bottom the larger the particles in motion will be, and the faster they will move. It has been found by observation that the bottom of the river is usually formed into a series of wave-like ridges composed of this movable material, the up

stream slope of these ridges being gentle and the downstream slop steep. Along the upstream slope the particles are moved by th current until they reach the crest, when they drop over. Particles are constantly being torn from the upstream slope and are rolled to and deposited on the downstream side of the crest, and the ridges. thus move steadily downstream.

The study of these ridges or sand waves is very difficult, for they are hidden from sight beneath the water of the river, which is so muddy that even an inch of it can not be seen through. The only places where these sand ridges can be actually seen are on bars submerged at high water and exposed in low water. After a fall of the river, the ridges on these bars become visible and are found of all sizes up to 10 feet or more in height. But even such observation is unsatisfactory, because it is impossible to tell at what stage and under what conditions the observed ridges were formed, and how they moved, and even if these points could be determined, such observations would still be of minor importance, for the waves thus examined were not formed in the main channel where the depth is greatest and the velocity swiftest, and it is in such swift currents and in the main channel of the river that the size of these waves is greatest and their motion most rapid and that the study of them is most important. Such ridges can only be studied by soundings, and the mapping out of the crests of these waves, the determination of their vertical sections and rate of motion where the water is deep and the current swift, is a difficult problem. Many observations of such waves have been made and a great deal has been learned about them, but in many respects the conclusions reached by different observers are contradictory, so that much still remains to be learned before the laws governing the formation and motion of these waves can be stated.

It is known that the shape and size of these waves, the distance apart of the crest of the waves, and their rate of motion depend on the depth of water and the velocity of the current, and that these functions of the waves are most regular when the stage and velocity of the current remain stationary for some time. A change in stage or of velocity conditions causes an entire or partial obliteration of the old waves and the formation of a new set suited to the new conditions. The waves are largest and farthest apart in the deepest water and the swiftest current, and it is in such conditions that their motion is the most rapid. Sand waves have been measured in the river more than 20 feet in height, with a rate of travel of as