Table 3 gives the progress of forest destruction by basins. It shows the number of square miles of original forest cleared and devoted to farm pasture or other uses and the percentage of the original forest area of the basin which this area represents:

The estimate of a total of 244,000 square miles of forest lands in the Mississippi Basin includes not only the commercial forests, farm woodlands, and other timberlands, but also includes about 35,000 miles of so-called "idle " or square 66 waste" lands. These are areas once timbered or adapted to timber growing which have been so denuded by overcutting and repeated fires that they are no longer capable of restocking themselves with valuable timber in commercial quantities. Hence they are of such low value that no one will make an effort to protect them. Lacking timber cover and continually robbed by fire of surface protection, they offer only low resistance to erosion.

66

About 10,000 square miles of such "idle" or waste" land is included within the total of 115,000 square miles of woodlands belonging in farm ownership. The farm woodlands are chiefly in regions. devoted primarily to agriculture.

The 129,000 square miles of forest land not in farm wood lots arein the regions where soil and topography generally are more favorable for forest growth than for farm use. Unfortunately, only about 104,000 square miles are occupied by tree growth of commercial species, 25,000 square miles falling in the "idle" or "waste" class.

HOW THE WATERSHED FORESTS WERE STUDIED

It was to the 244,000 square miles of forest lands remaining on the Mississippi watershed, some timbered and some idle or waste, scattered through the 6 major basins and 73 individual drainage units, that this study was directed. It is desirable to make clear at this point that the value of a forest in preventing erosion and regulating stream flow is not directly proportional to the volume of timber on any given It is influenced by other factors, such as density of cover, amount and character of undergrowth, and depth of humus accumu-

area.

lation. A mature forest in open stand may support a large amount of merchantable saw timber and yet, because of overgrazing or repeated ground fires, be so devoid of undergrowth and leaf litter as to be rated very low from a protection standpoint. On the other hand, a watershed may be very well protected by a young or scrubby growth or by thickets of inferior species by reason of the density of its canopy and the depth of litter accumulation.

Previous general studies of the forests of different regions of the United States have been made principally for the purpose of determining the value of the forests as sources of lumber or other wood products or of estimating their timber-producing possibilities. But the flood problem required a study of forest areas markedly differing from anything ever before undertaken on so large a scale. The primary purpose was to find out, if possible, to what extent our forests may be relied upon to help reduce floods and flood damage on the Mississippi.

This required at the very outset a study of the forests and forest lands of the basin to determine the extent and location of areas having important protective significance as distinguished from areas from which run-off was not materially influenced by the presence or absence of forest cover.

In the plan of this study an endeavor was made first to localize the source of the watershed trouble so far as was found practicable, in order that effort might not be wasted, and that attention might be focused on the "critical areas" rather than diverted to others of less importance.

To be of any substantial or permanent value, it was evident that the classification of the areas could not be based entirely upon the immediate efficiency of each watershed. It was obviously just as important in the long run to prevent a good-protection forest from being injured as to restore the efficiency of a similar area which had already been injured. Just as important and much cheaper and easier.

There was therefore placed in the "critical areas" class all forest lands found to be of such character, by reason of factors of soil, slope, and precipitation, that to prevent flashy run-off or destructive erosion the maintenance of a good forest cover was necessary as proved by the results of forest destruction on similar land. It was upon the areas so classified as "critical areas" that this study finally centered. Under this general designation of "critical areas two general regions not falling in the class of forest lands were finally included. These are the Bad Lands and the Breaks.

How determined.-The determination of "critical areas" involved the consideration of many complex and conflicting factors. While the three main factors which determined the final classification (soil characteristics, physiography, and precipitation) are relatively con

3 See fig. 6 and Table 7.

stant for each part of a drainage unit, they vary for different parts of the unit. Furthermore, each factor is exceedingly complex. For example, an average precipitation of 40 inches annually in the form of an occasional torrential rain will have a very different effect from the same total amount of precipitation in many gentle rains well distributed throughout the year. Any system of measurement based upon factors so variable is obviously not infallible. At the same time it has furnished the Forest Service with a substantial background of information which, when checked by run-off conditions on the ground, is reasonably accurate. At least it gives different examiners in different regions common standards of measurement and

make comparisons possible. Faulty as this method may be, wide diversity of conditions required the adoption of some uniform method of rating watersheds in order to make possible the summations and comparisons absolutely essential to any real conception of actual conditions on such a large area.

By means of the rating system adopted Figure 6 now gives for thefirst time a graphic presentation of the areas where the maintenance of good forest cover is believed to have an important relationship to flood control.

The location of "critical areas" having been determined, a further classification was made according to present conditions. Those which were not the source of heavy deposits of silt or flashy run-off were