land. (See fig. 4.) Comparatively speaking, it has always been a lightly forested country. Originally 60 per cent, it is now 80 per cent non forest land.

Also floods in the Mississippi have been recorded as far back as history reaches. De la Vega, in describing the difficulties encountered by De Soto's men when, following their leader's death, they determined to build boats and descend the Mississippi, says:

Then God, our Lord, hindered the work with a mighty flood of the great river, which, at this time, began to come down with an enormous increase of water, which in the beginning overflowed the wide level ground between the river and the cliffs; then little by little it rose to the top of the cliffs. Soon it began to flow over the fields in an immense flood, and as the land was level, without any hills, there was nothing to stop the inundation. The flood was 40 days in reaching its greatest height, which was the 20th of April, and it was a beautiful thing to look upon the sea where there had been fields, for on each side of the river the water extended over 20 leagues of land, and all of this area was navigated by canoes, and nothing was seen but the tops of the tallest trees.

The report on this river submitted in 1861 to the Bureau of Topographic Engineers, War Department, by A. A. Humphreys and H. L. Abbott lists floods as having occurred in 1718, 1735, 1770, 1782, 1785, 1791, 1796, 1799, 1809, 1811, 1813, 1815, 1916, 1823, 1824, 1828, 1838, 1844, 1847, 1849, 1850, 1851, 1858, and 1859. The serious floods of more recent times have been in 1882, 1892, 1897, 1898, 1903, 1907, 1912, 1913, 1916, 1920, 1922, and 1927. The highest recorded stage of water previous to 1927 on the lower portions of the river, was reached by the flood of 1922.

It is difficult to make a fair comparison as to severity between the later floods and those of earlier times. Before the construction of levees high water every year inundated a considerable area of low country. Naturally, as the levee system has been constructed and the stream confined, the tendency has been for the waters to reach higher and higher levels. Even without any increase in precipitation or run-off, this would be the natural result of excluding the flood waters from lands formerly subject to overflow.

It was early evident that for reliable information on how run-off was being influenced by the presence or absence of forests, a study must be made in detail of the actual source of run-off. The regimen of a stream at any given point is a composite of all factors influencing run-off from the entire area above. Influences constructively beneficial on one drainage unit may be neutralized by destructive factors on some other contributing stream. Obviously the flood waters as they reach the leveed region are a mixture coming from all kinds of watersheds, some well protected, some poorly, some not at all.

The total area of the Mississippi River Basin within the United States, including the lands subject to inundation, is 1,231,492 square miles. So vast an area is hard to treat as a unit. Diversity of climate, cover, and human use adds complexity to immensity. Although only about 244,000 square miles of forest lands, or about 20 per cent of the total area, are involved in the problem, these forest lands are distributed throughout the humid portions of the entire basin. Furthermore, the use of forest lands must be coordinated on a practical basis with the use of other kinds of lands adjoining.

Though the source of the great volume of water in the 1927 flood was largely the run-off from heavy and prolonged rainfall in the lower

central valley, at the same time a factor to be reckoned with at all times is the contribution which each individual drainage unit may reasonably be expected to make to the main flow at any given time. For a thorough understanding of the problem it was therefore necessary to consider the watershed both in detail and as a unit.

IMPORTANT DATA LACKING

In the course of this study it was found that reliable data essential to accurate determination were in many instances nonexistent. Topographic maps, soil maps, forest maps were too often lacking for important regions.

It seems fundamental to a carefully planned study of land use, covering long-time periods, that there be available accurate data for each of the major tributaries of the river showing annual and periodic run-off and total and periodic volumes of silt content. Dependable data on silt content were woefully lacking. Determinations made many years ago could not be checked against similar data of more recent date because recent measurements had not been made. A complete record of stream-flow measurements and silt determinations at some point below Cairo and on each of the principal tributaries entering the Mississippi River below that point and covering the past four decades would have been of inestimable value in this study. Such records beginning in 1928 would become increasingly valuable from year to year. Regardless of any light the data might throw on the flood problem, it would give a splendid check on the problem of land use and the trend of soil erosion. By means of such records it would be possible at any time to determine whether or not efforts in reforestation, forest protection, contour plowing, and farm terracing were reducing the burden of silt in the river at Cairo or in any of the tributaries under observation. For that reason the absence of such data is noted in this report, and the desirability of securing accurate information of this kind is given a place in the conclusions and recommendations.

MAJOR BASINS OF THE MISSISSIPPI

A systematic investigation of the forested watersheds of the Mississippi Valley and their relation to the floods of that river required a somewhat arbitrary subdivision of the entire region, first into major basins and then into individual drainage units.

The subdivision into six major basins which was finally adopted for this study is shown in Figure 2. This subdivision appears to be generally followed, with perhaps some minor modifications, by all agencies in studying the Mississippi flood problem. It follows largely the classification used in the report of 1861 by Humphreys and Abbott. These authors, however, treated the St. Francis River and the Yazoo Basin each as distinct major basins. The subdivision of the major basins into principal drainage units used by the Forest Service in making the study, with the areas of each major basin and each drainage unit thereunder, is shown in Table 1.

TABLE 1.—Mississippi River drainage—Areas of major and minor basins within the United States'

1 Based on planimetric measurements of U. S. Geological Survey maps applied to the actual surveyed areas of the States. Canadian drainage of 12,934 square miles in the Missouri Basin is not included. The drainage of the Mississippi River proper, excluding the Ohio, Missouri, Arkansas, and Red River Basins, is divided at Cairo, Ill., into upper Mississippi and lower Mississippi.

* Includes 5,441 square miles of indeterminate drainage such as the Atchafalaya River and other interconnected waterways in southern Louisiana.

For a proper understanding of the protective influence of the forests of the Mississippi River drainage, their relationship to floods, and their possibilities and potentialities as factors in stream-flow regulation, due consideration must be given to certain well-recognized characteristics of running water and the relationship which

19