stages the bed fills up in these shallow places, so the deeper water and higher velocity instead of giving a lower bed to the river gives a higher bed to the river in very wide places, which are the shallow ones. Mr. BRICKNER. Does it not create more pressure the more water you get in it? General COMSTOCK. There is more weight of water on the shoal. Mr. BRICKNER. After the velocity is increased does that collect sediment on the bottom? General COMSTOCK. Yes. On this Plum Point reach, where it is 14 miles wide, I said they increased all the way up. That is not strictly accurate, because much of these bars is formed when the water is falling. Mr. BRICKNER. That is true; but suppose that it narrowed in this continuous levee and this continuous levee would hold the water closely together, would it become as low or as shallow as it will by throwing out these spurs and allowing— General COMSTOCK. I do not think I quite catch your question. Mr. BRICKNER. Would these bars form if there was a continuous levee ? General COMSTOCK. They form now at wide places when the river rises 30 and 40 feet, and I think they would form then. I do not think it would change the action on the bed, and I do not think it would ent the bed out, but it is impossible to say just exactly the importance of these facts. I only judge from the fact that the bed itself rises. Mr. BRICKNER. Now, I understand these spurs are put out diagonally, like this [illustrating with a book]. General COMSTOCK. Very often at right angles, square to the bank. Mr. BRICKNER. Does it not break over it in high water? General COMSTOCK. They are pile dikes of two, three, or four rows, and about 10 feet apart in the rows, and sometimes brush is woven in making a kind of wicker work. The water flows through and the result of that is that the sand is deposited right below. Mr. BOATNER. Is it not an invariable result that there is an immediate accumulation of silt behind these dikes, I mean these spur-dikes, unless there is a very strong current there? General COMSTOCK. That is what we build them for, and we almost always get an accumulation of silt behind them. Mr. BOATNER. That proves conclusively that any interruption to the current of the Mississippi River causes an immediate precipitation and deposit. General COMSTOCK. No; it proves that the interruption which we form by those dikes cause a precipitation and deposit. Mr. BOATNER. In the course of your investigation have you found any interruption to the current of the Mississippi River which did not cause precipitation and deposit of silt carried in the current ? General COMSTOCK. That is a question which is difficult to answer. That seems to imply that the higher the velocity the more the sediment carried. That is not true; it is very far from true. Mr. BOATNER. This is true, that the carrying capacity of the river or the capacity of the water to maintain and to carry is in the ratio of the velocity of the current. General COMSTOCK. What its carrying capacity is I do not know. I know what it carries, and I know it often carries much more sediment at medium velocities than at the highest; it depends not only upon the velocity of the water but on the fineness and supply of sediment, and the supply of sediment comes largely from the Missouri River, is very fine, and is carried a long way. The amount of sediment a river carries depends upon the velocity and also upon its fineness. Gravel the Mississippi River will scarcely carry at all, whereas clay is carried from Cairo to the Gulf of Mexico by low velocities as well as high. Mr. BRICKNER. I was raised there on that bottom and I know by actual experience and observation that before the mere interruption of fence, in the case of an overflow where a fence does not wash away, or there is a piece of levee-say there is a crevasse here and a crevasse there [illustrating], that there is invariably a large deposit immediately behind the standing levee or immediately behind the fence, and I have never known it to fail. Now, I will ask you as a gentleman who understands these subjects, if that is not conclusive evidence of the fact that the interruption of the current which causes temporary stagnation of the water does not cause a deposit, and that is what causes the formation of sand-bars and other bars in the bed of the river? General COMSTOCK. I do not doubt that large crevasses form sandbars, and in the bed of the river in some places, but the question is not whether they form bars in general, but whether they build up low-water bars, which impede navigation. It is undoubtedly a fact that during high stages these bars build themselves up and are cut down by the low stage of the river. I make a distinction, which is, that where you have deep water and a tolerably narrow place, if you increase the height of the water by levees the probability is you will scour out the bed of the river there, and if you diminish the height of the water you may cause a partial deposit; but the worst bars are built up higher during high stages and are cut down when the river falls. Mr. BOATNER. Are the banks on Plum Point Reach high? General COMSTOCK. They are moderate banks; they are not as high as Vicksburg or Natchez. The river oscillation is only about 35 feet. Mr. BOATNER. And have the levees been well maintained on that reach for the last 20 years? General COMSTOCK. NO. The levees on the St. Francis front are in a very bad condition as a rule. The Commission has built a line of levees at Plum Point Reach, and some private owners there, I think, extended them somewhat on the west bank. On the west bank there is a good line of levees, and I think it must be 20 or 30 miles long. Mr. BOATNER. How long since that levee has been completed and made a continuous line? General COMSTOCK. I think that work was done about 1888. Mr. BOATNER. Have you observed any improvement in the depth of the water since the construction of this levee ? General COMSTOCK. I am not able to say. I think prior to the construction of this levee in 1887 the river went down, I think, to 8 feet, in 1886, and since that levee was built, one year, 1887, I think the river went to 8 feet, and in 1889 while the least depth was 10 feet the river was not as low by 1 foot as in 1887. That is to say, we built contraction works producing an increase in depth. Then we built the levee, and now it is very difficult to separate the effect of the levee from that of the contraction work. My judgment is if the levee had not been built you would have about as much water as now. Mr. BOATNER. You do not concur in the theory that the confinement of water between levees increases the velocity of the current, and thereby increases the capacity of the river to carry sediment and also to scour out the bottom and increase the depth? General COMSTOCK. Some of it I agree to and some I do not. More water increases the velocity of a river, but on bars inasmuch as when the river rises say 20 to 40 feet and then falls again, these bars are built up I am not sure the building of levees is going to increase the depth on the bars. The CHAIRMAN. Your judgment is the scouring of the bed of the river, except where there is very deep water, and the cutting of the bars occurs when the water receded, and it is reached by channel works instead of high velocity? General COMSTOCK. Yes, sir; the lowering of the bar is done during the low-water stage and the raising is done during the high-water stage. Mr. STOCKDALE. What conditions, if you have any opinion on that subject, produces these bars? Can you attribute it to the current or form of the soil, or what in your opinion is the cause of these bars being at these particular points? General COMSTOCK. That question has not come up in the Commission, but I think I express the general views of the Commission when I say that the cause of these bars is the width of the river. At one time some of the members of the Commission thought there was gravel in the river which caused the bars, but I do not think there is. I think it is merely the width of the river. The river on a straight reach wanders about, and if it happens to take a notion now to scour into one bank and after that the other bank, it does it, and that widens the river. Mr. STOCKDALE. How would it do say, to produce these dikes from both banks General COMSTO K. That is exactly what we do. It is this way [illustrating]: We do not do it from both because one bank has deeper water than the other and it is better to build it up in the shallow, so we generally put our dikes out from the convex side. That is what we did at Plum Point. In some cases it may be necessary to put the dikes on both sides. Mr. STOCKDALE. Of course the banks of the river are lower opposite these bars or shallows than they are elsewhere; that is, they are lower in comparison with the bed of the river. General COMSTOCK. Yes, sir. Mr. STOCKDALE. That is the very point I wanted to ask; if these levees were built on these banks where you have your dike to keep the water from scouring out these places between? General COMSTOCK. As I said before, I think we have some experience which does not yet show clearly the advantage of the levees, and then this question of the raising of the bars that I speak of— Mr. STOCKDALE. I have always heard that the bars rise during high water. For instance, if the water is at one stage, say 20 feet, when the water gets to 30 it will not be that much deeper, but the bottom will rise 5 feet where the surface rises 10, Mr. BOATNER. Is it universal that the low banks are found opposite these shallow places in the river? General COMSTOCK. No; the height of the bank, in my opinion, is determined very largely by the amount of water which is flowing in the river. Mr. BOATNER. What I judge by is the depth of the country overflowed. There are some points you need a levee 12 feet, and other points where you need a levee 2 feet, and there are some points where no levee at all is necessary; but I mean to say, is not the bank relatively low wherever you find these shallow places? That is, would not it require a high levee to protect the country from floods opposite these shallow places in the river? General COMSTOCK. No; there are shallow places where the banks are high. The height of the bank is determined by the amount of water flowing in the river. At Cairo, in 1882, there was 1,800,000 cubic feet of water flowing per second in the river, and the river rose 52 feet above low water. Going down to Plum Point Reach there was a flow, in the flood of 1882, of about 1,200,000 cubic feet per second, which is 600,000 cubic feet per second less than at Cairo, and high water was about 35 feet above low water. At Lake Providence Front the river flood is something over 40 feet, and the banks are low. At Vicksburg it is over 50 feet, and the banks are high. At Lake Providence you had, in 1882, only about 1,000,000 cubic feet per second flowing there, and at Vicksburg you probably had 1,500,000 cubic feet per second. Mr. BOATNER. Is it your recollection that the banks are lower at Lake Providence than opposite Vicksburg? General COMSTOCK. I do not speak of any definite point. I only speak about them in a general way. This oscillation is what fixes the height of the bank. Mr. BOATNER. I am afraid I have not made myself understood. You know the bank of the Mississippi rises aud falls. General COMSTOCK. Yes, sir. Mr. BOATNER. There are places where the river has cut through bends and the bank is high, and there are other places where the lowland runs out nearly to the river, slopes up very gradually, and where the water from time immemorial has run out. I call your attention to the fact that there are a number of bayous which formerly connected the Ouachita system and the Mississippi River, and where they formerly put out from the river the banks are low. Now, what I want to get from you is this: Is not it a fact that the banks of the river are low opposite to these shallow places to which you refer? General COMSTOCK. You mean low counted from low water? Mr. BOATNER. I mean taking the highest point on the river as a starting point. General COMSTOCK. That is a point I have not examined, and I am not prepared to say that as compared with other banks they are low. Mr. STOCKDALE. I do not exactly understand that the fluctuation of the river has decreased uniformly as it goes towards the mouth, but I understand the fluctuation is greater at Vicksburg than at some points above there. General COMSTOCK. Yes, sir. Mr. STOCKDALE. How do you account for that, by overflows? General COMSTOCK. Yes, sir; when you get down to Helena the fluctuation there is [examining memorandum] Cairo extreme oscillation is 53 feet, Plum Point, 34.5, and the cause of it is that enormous escape along the St. Francis Front. At Helena the extreme oscillation is 48 feet. At Vicksburg the extreme oscillation is about 53 feet, 10 feet more than Lake Providence only a few miles above, and the cause of that at Vicksburg is the enormous return flow of the Yazoo bottom which raises the river there and increases the oscillation. So the amount of oscillation depends upon the amount of water flowing in the main river. Where you have the largest amount, something like 1,600,000 feet per second, which is about the largest amount anywhere in the river now, you have an oscillation of from 40 to 45 feet and more. Mr. STOCKDALE. You make that computation at what stage of the water, the highest? General COMSTOCK. Yes, sir. That was the flood of 1882 which measured 1,600,000 at Columbus, nearly the same amount at Helena, and 1,500,000 at the Red River. Mr. STOCKDALE. How far do you go, just to the bank; suppose the water is over the country? General COMSTOCK. In these figures I confine myself to the river alone. We measure also that out in the country. Mr. BRICKNER. As I understand, the bars form only when the water is high, at the highest point? General COMSTOCK During the high stages of water where you have, say, a wide and shallow place with 6 feet at low water, as a rule the bar rises during the high stages of the river, and when the river falls and gets down to 10 or 15 feet above low water instead of having 15 plus 6, which would be 21 feet of water, you may have only 15 feet. Mr. BRICKNER. That is because the velocity of the water being decreased as it widened out into the country. General COMSTOCK. I think the cause is that the velocity of that wide place is considerably less than it is in the deeper places above it, the deep and narrow places above and below, and that the bed has risen during high water. Mr. BRICKNER. If that could be confined and the velocity of the water could be equaled above and below, would it then form bars! General COMSTOCK. If you narrow it down to the same width as above and give more depth, then the bars would disappear in part— would be lowered-that is what we are trying to do by the contraction works. Mr. BRICKNER. That is what I could not understand. Mr. BOATNER. Have you any record of the number of miles of levee which gave way in the flood of 182? General COMSTOCK. Nothing very accurate. Mr. BOATNER. Have you a record of what gave way in 1890? Mr. BOATNER. Is not it less than those breaks in 1882? General COMSTOCK. Yes; it is very much less than 1882, and in 1882 also were old breaks which occured in the 10 or 20 years preceding; there were very large gaps, I think maybe 60 miles of gaps in 1882. Mr. BOATNER. Was there any line of levee from St. Francis Front to the mouth of the Red River, practically; was not the water running practically broadcast over the country in 1882 ? General COMSTOCK. They were practically down as far as Arkansas City, and then from Arkansas City down to the Red River they were very badly broken in 1882 on both banks of the river. Mr. BOATNER. The flow of the water was practically uninterrupted from the west bank of the river to the Ouachita system, was it not? General COMSTOCK. I am not able to say whether one-third of the levees were down or what percentage it was, as I do not recollect it this moment. Mr. BOATNER. Can you state what was the difference in the flood height of the water between Vicksburg and the mouth of the Red River at the high stage of the water in 1882 and 1890? General COMSTOCK (consulting memoranda). Lake Providence, 2.7; Vicksburg, 0.3; Red River-I have not got a note of it here. Mr. BOATNER. How much higher in 1882 than in 1890 ? General COMSTOCK. It was 2.7 feet higher at Lake Providence in 1890, and three-tenths at Vicksburg. Mr. BOATNER. At both of those places the flood height was actually higher in 1890 than in 1882 ? General COMSTOCK. Yes, sir. Mr. BOATNER. As a fact, in 1882, the whole country west between |