years, unless an immense annual outlay were made, the country would probably be as much exposed to overflow as at present. The Arkansas levee district, included between Helena and White Rivers, is more fortunate than that just considered, in being perfectly protected from overflow from above by the Helena hills. It is, however, very small in extent. From Helena to Oldtown Ridge there were, in 1866, five breaks, contents 148,000 cubic yards, costing about $59,000 for repairs. From Oldtown Ridge to Carson's Landing, near Islands 67 and 68, there were then several breaks, the worst being at the Lima place total contents about 150,000 cubic yards, requiring about $60,000. Thence to Laconia, the levees were good, (distance fifteen miles.) At Laconia the planters had repaired the State levee to Bob's Bayou, which enters the Mississippi two miles below Island 71, a distance of seven miles, and had connected these termini by a rear levee eleven miles long, to keep out the overflow from White River, thus inclosing 15,000 acres of good land. The Laconia circuit broke in the upper part in 1867, and had to be cut below to let out the water. From Bob's Bayou to Napoleon no levees ever existed. Probably the cost of repairs in this entire district would not exceed $150,000; but its protection is strictly a local matter. The third Arkansas district, or rather that part of it lying below Gaines's Landing, has already been considered in discussing the protection of Northern Louisiana. Above Gaines's Landing, to Desha County line, the levee is reported good. Thence to Napoleon it is practically gone. To protect this region from floods in Arkansas River, and from Mississippi back-water in that stream, levees have been extended up the river from near Napoleon to a point forty-five miles below Little Rock, chiefly on the southern bank. The upper part of these levees is reported good, but several breaks are named, particularly at four, at ten, and at fifteen miles above Napoleon, and near Heckatoo plantation. To close the breaks in both rivers sufficiently to protect the region above Gaines's Landing from overflow, (present grade,) would probably involve an outlay of $300,000, making the total cost of repairing breaks in this third levee district of the State of Arkansas about $700,000. To close the breaks necessary to protect the entire State would then call for an expenditure of $1,850,000. In fine, then, to close the breaks now existing in the Mississippi levees, would cost as follows: State of Louisiana. State of Arkansas. $1,050, 000 1, 500, 000 1, 850,000 It should be remembered, however, that in these estimates the cost of excluding crevasse water, entering Louisiana through Chicot County, Arkansas, ($250,000,) is contained twice, once for Arkansas and once for Louisiana, and that the outlay required for Arkansas also gives protection to a part of Missouri. It should also be borne in mind that this money is required to simply replace the levee system of the Mississippi where it has already been when most complete. This is far below what the real security of the region demands, and the only justification for such an expenditure would be found in the fact that it would enable planters to make crops in ordinary seasons, well knowing that at any recurrence of great floods, which happens usually about once in three or four years, extensive inundations would be sure to occur. In the suffering and reduced condition of the region at present, some such temporary and partial relief might enable the planters to obtain enough funds and credit to save their estates 335 from ruin, and thus prepare, eventually, for building the more extensive levees which security demands. A PERFECTED LEVEE SYSTEM. The investigations and surveys conducted by yourself between the years 1850 and 1861, and fully elaborated in the report upon the physics and hydraulics of the Mississippi, which constitutes professional papers No. 13 of the corps of engineers, have demonstrated that the best, and, indeed, the only feasible method of protecting the alluvial region from overflow, is that of a levee system in which the dimensions of the embankment are computed to restrain the maximum flood discharge of the river when confined to the channel from Cape Girardeau to the mouth. In that report the whole subject is thoroughly discussed, and the dimensions of the levees in all parts of the region are computed in detail from the very elaborate and exact data obtained by actual measurement in the flood of 1858. That flood was adopted as the standard, because a close comparative analysis of all other recorded floods, including that of 1859, proved that in no other would the maximum discharge have been in excess of what would have occurred in that flood had the levees been able to restrain the river to its bed. Hence, at the date of that report (1861) the probable difficulty and cost of a perfected system which should give to the plantations upon the banks of the Mississippi the same security that is enjoyed by the fields of Holland, was accurately known. The only point which demanded further investigation was, whether the flood of 1858 had been correctly assumed as a standard, a point which time alone could certainly determine. Since 1859, there have been but three great flood years-1862, 1865, and 1867; the others belong to the class of ordinary high waters, in which the projected levees would have largely exceeded the requirements of the maximum volume. To decide, therefore, at the present time upon the proper dimensions of levees for the Mississippi, we have only to compare carefully those three great floods with that of 1858, to ascertain whether or not the water-marks and recorded facts indicate a maximum discharge at the head of the alluvial region, or just below the mouths of any of the lower tributaries, in excess of that which would have occurred in 1858, had all the water been confined to the channel from Cape Girardeau to the Gulf. If this question be decided in the negative, the flood of 1858 remains a safe standard; if in the affirmative, the estimates in the physics and hydraulics of the Mississippi must be modified to allow for the increased volume to be apprehended. The first point, then, for attention is, the extent of the information which has been preserved respecting the three great floods in question. When acting as your assistant, upon the examination of the levees, in the winter of 1865-66, I made every effort to collect all possible facts respecting the floods of 1862 and 1865. Sufficient high-water marks were found to indicate, with a good deal of precision, the level attained by each of these floods, as compared with that of 1858, throughout the alluvial region. Through the kindness of Mr. Aug. V. Taylor, at Cairo, and of Mr. G. W. R. Bayley, at New Orleans, daily records of the stand of the river at those points in 1865 were received. Some meagre information respecting the condition of the different tributaries during the two floods was also secured; but the war had distracted attention from river phenomena, and the lapse of time had rendered it impossible to collect as full data as could be desired, especially for the flood of 1862. Before the flood of 1867 had subsided, instructions were issued from the headquarters of the Corps of Engineers to Brevet Brigadier General McAlester, at New Orleans, to Brevet Colonel Merrill, at St. Louis, to Brevet Major Burroughs, at Nashville, and to Mr. W. Milnor Roberts, superintending engineer of Ohio River improvements, to collect all possible data respecting the overflow. Circular letters were accorddingly at once addressed to the different civil and military authorities, requesting facts. Many valuable letters were received in reply. This material accompanied your instructions directing me to prepare this report, and upon it, and a few other data received from Mr. S. Staats Taylor, at Cairo, and from Colonel Merrill, at St. Louis, the following analysis of this flood is based. Before proceeding to the detailed discussion of the three floods, the following table is presented to exhibit their relative high-water marks, as compared with the floods of 1858 and 1859. It is properly a continuation of the flood-table on page 170 of Physics and Hydraulics of the Mississippi; but the flood level of 1862 has necessarily been adopted as the plane of reference, instead of that of 1858. The sign + denotes that the flood in question exceeded the height attained in 1862, and the sign that it fell short of that height. The numbers following the signs denote the difference in height attained in the two floods, expressed in feet. In comparing the high-water levels in these different floods, the fact must be borne in mind that four cut-offs have occurred during the period, viz: the American Bend cut-off, on April 15, 1858; the Napoleon cut-off, on April 11, 1863; the Terrapin Neck cut-off, early in March, 1866; and the Davis cut-off, at Palmyra Bend, on February 10, 1867. Their relative positions are indicated in the table. Flood of 1867.-As the records of this flood are more complete than of either of the others, it will be considered first. In some respects its origin was peculiar. The winter of 1866-67 was marked throughout the southern portion of the Ohio Valley by an unusual down-fall of snow and rain; while in the region drained by the Upper Mississippi and Lower Missouri the season was remarkably dry.. A sudden thaw with warm rains in February caused moderate floods in the Alleghany and Monongahela rivers, and in the smaller tributaries of the Ohio, leading near the main stream; and a great flood, second only to the flood of 1858, in the Wabash. The combined effects of those freshets was to cause a very sudden rise in the Ohio, which culminated at Louisville on February 22, where it was only eight feet below the high water of 1832; and at Caseyville, below the mouth of the Wabash, on March 1, where it was half a foot above the high water of 1832, the greatest of the recorded floods at that locality. The same climatic influences extended over the valleys of the Illinois. River and other southern tributaries of the Upper Mississippi, producing a moderate freshet in the Mississippi at St. Louis. The rise there began on February 13, the river being 25.5 feet below the city directrix; it culminated on February 21, at 9.3 feet below this bench; after remaining four days sensibly at a stand, the river gradually subsided, until, on March 21, it was 21 feet below the directrix. The freshet at St. Louis was by no means a large one, being 16.9 feet below the high water of 1844, and 12.6 feet below that of 1858; still, it is evident that it almost exactly combined at Cairo with the February rise in the Ohio, and thus. did its maximum of injury to the alluvial regions. The downfall at St. Louis was 2.3 inches in January, 4.8 inches in February, and 2.4 inches in March, showing a slight indication of the great February rains, but none whatever of those in March. Such was the condition of the rivers, when, in March, a wide-spread series of furious rain-storms occurred. The belt containing them extended from the headwaters of the Washita and White Rivers of Arkansas, eastward across the States of Arkansas, Missouri, Kentucky, Tennessee, Western North Carolina, and Western Virginia; but it was in the mountain region, where heads the Tennessee River, that the greatest deluge occurred. The downfall here was entirely beyond precedent, raising the Tennessee River at Chattanooga, on March 11, 53 feet above low water; or 15.5 feet above any known water-mark. With the Cumberland, the Kentucky, the Green, and, indeed, all the lower southern tributaries discharging full floods into the Ohio, before the February rise had had time to pass away, this sudden Tennessee River flood raised the Lower Ohio to the highest stand ever attained. Fortunately the immense wave found the Mississippi burdened only with the previous rise, the Upper Mississippi, the Missouri, and the Arkansas all being low. The Washita, White, St. Francis, and Yazoo Rivers were swollen from the same rains, but probably not sufficiently to produce much effect upon the great wave from the Ohio, which arrived rather too late to coincide with their freshets. This flood in the Ohio was no less remarkable for duration than for extreme height-matters of equal importance in effecting a flood in an immense channel like that of the Lower Mississippi. For 32 consecutive days, at Cincinnati, (February 16 to March 19,) the mean channel depth was 51.3 feet, the greatest being 55.8 feet and 57.3 feet, on February 22, and March 14 and 15, respectively, and the least being 44.6 feet, on March 2 and 3. So long a continuance at this stage is beyond precedent. These facts respecting the tributaries make it evident that the flood of 1867 in the Mississippi itself must have been marked by many peculiarities. In order to convey a clear idea of its character I have prepared the accompanying plate, drawn upon the same scale as that adopted in the Report upon the Physics and Hydraulics of the Mississippi, to illustrate the floods there discussed. [For plate referred to, see original on file in office of Chief of Engineers.] It will be noticed that, beside the oscillations at various points in 1867 and 1865, the plate contains other valuable river records, obtained through the kindness of Mr. Taylor and Mr. Bayley. Since there was no great Mississippi flood above the mouth of the Ohio in 1867, Cairo is the first point which requires attention. It is a particularly important locality in all floods, being situated so near the head of the alluvial region that, when the source of the flood is known, a relative estimate of the maximum discharge into that district may be formed from a judicious study of the gauge indications there; but in this connection it is well to call attention to the following facts, which were fully established by repeated observations upon the Mississippi, and which, paradoxical as they may appear, are in perfect accordance with the laws governing flowing water.-(See page 324 Physics and Hydraulics of the Mississippi.) 1. For any given stand there is much more water passing when the river is rising than when it is falling. 2. For any given stand there is usually more water passing in a long and rapid than in a short and slow rise; but this is not always .the case, the discharge being governed by the relative stage of the water in the channel above and below. 3. The maximum discharge in any normal rise occurs when the river has reached a point a few inches below the highest point attained. 4. If, when a freshet has culminated, and the water either comes to a stand or begins to fall, a second rise occurs, it will cause the surface to rise considerably higher than would have been the case had the same volume passed without a previous diminution of supply. For instance, in the flood of 1851, the Mississippi at Red River Landing attained a certain stage, with a measured discharge of 1,200,000 cubic feet per second. It had ceased to rise, and was just ready to begin to fall, with |
