The CHAIRMAN. At any rate, none of these letters were published by the Forest Service, to your knowledge?

Captain JOHNSTON. No, sir.

The CHAIRMAN. Are there any further questions? If not, we are very much obliged to you, Captain Johnston.

(The committee thereupon adjourned until to-morrow, Thursday, March 3, 1910.)

(The letter of March 3, 1909, from Maj. William H. Harts, Corps of Engineers, U. S. Army, to the Chief of Engineers, U. S. Army, read in part by Captain Johnston, and directed to be incorporated in full in the record, is as follows:)

WAR DEPARTMENT,

ENGINEER'S OFFICE, UNITED STATES ARMY, Custom-House, Nashville, Tenn., March 3, 1909.

The CHIEF OF ENGINEERS, U. S. ARMY,

Washington, D. C.

GENERAL: I have the honor to submit the following reply to your letter of November 5, 1908 (E. D. 68551), referring to the general subject of the influence of forests upon stream flow.

1. The main rivers in my district are the Cumberland and the Tennessee. These streams are somewhat similar in their upper regions. They take their source in similar country, the southern Appalachian region, and follow the same general direction of flow. Their drainage basins adjoin; that of the Tennessee lying to the south.

The tributaries of the Tennessee extend farther eastward and over a much wider range of latitude; the area drained by them being approximately double that of the Cumberland basin. The hydrographic records of these two streams, as far as they have been kept, seem to point out plainly the fallacy of the claim that deforestation is noticeably injuring the navigable capacity of our rivers.

2. The Cumberland with its tributaries has its sources in eastern Kentucky and Tennessee at elevations of about 2,000 feet above sea level. The drainage area above Burnside, Ky., is 3,739 square miles and above Nashville 11,600 square miles. As is to be expected, the slope of the river and the general grade of the country through which it flows are much steeper in the upper than in the lower portions, the average fall of the river surface from the mouth of Poor Fork to the mouth of Rockcastle River being 3.04 feet per mile; from the mouth of Rockcastle River to Burnside 2.57 feet; from Burnside to Nashville 0.67 foot; and from Nashville to the mouth 0.403 foot.

3. The Tennessee is formed near Knoxville by the confluence of the Holston and French Broad rivers. These tributaries rise in western Virginia and North Carolina at elevations of 2,500 to 3,000 feet above sea level. The drainage area above Knoxville is 10,295 square miles, of which 6,438 square miles belong to the French Broad and 3,857 to the Holston. The drainage area above Chattanooga is 21,418 square miles. The average fall of the Holston from Kingsport to the main river, a distance of 142 miles, is 2.54 feet per mile. The French Broad from Brevard, N. C., to Leadvale, Tenn., descends at an average rate of 8 feet per mile, while from Leadvale to the mouth its average fall per mile is 2.27 feet. From the confluence of these two rivers to Chattanooga (188 miles) the fall in the Tennessee averages 0.95 foot per mile.

4. The country in the vicinity of the upper portions of these river systems is generally wooded, the areas of cleared and cultivated land increasing as we pass down the streams. The river bottoms are often highly cultivated, but the soil of the uplands is generally poor, and owing to the mountainous character of the country surrounding the headwaters these regions are of far less value for agricultural purposes than the valley lands lower down. The areas of cultivated land are increasing, however, and vast areas of timber land have been cut over within the past fifteen years, so that deforestation in these areas is referred to by the Forestry Bureau as already dangerous in extent and as increasing with alarming rapidity. They maintain that it is now affecting adversely the navigable condition of our streams, and that forest preservation is necessary if we expect to protect their navigability from injury.

5. Exact information as to the areas under forest cover at the present time is very meager, and many efforts to obtain even a close approximation were fruitless, so that the estimates of forested areas given herein are largely based upon conjectures. From the best information available, I should say that the proportion of the drainage basins of these streams under forest cover would now be in the neighborhood of 60 per cent. The increase in cleared areas within the past twenty-five years, including tracts deforested for the sake of lumber, may perhaps amount to as much as 20 per cent. From the census report for 1900 it appears that in Tennessee the increase in the areas of improved farm lands between 1880 and 1900 amounted to 1.749.394 acres; in North Carolina to 1,845,915 acres, and in Virginia to 1,584,692 acres. The changes within twenty years in the lower lands of these States have not been nearly so extensive as in the higher portions, and it may be assumed without much question that the greater part of this increase was at the expense of the forested areas about the headwaters of the streams under examination. The total amount of clearing in the southern Appalachian region is given by the Forestry Bureau as 24 per cent (Report of the Secretary of Agriculture on the Forests, Rivers, and Mountains of the Southern Appalachian Region, 1902, p. 26), leaving an average of 76 per cent still under forest cover.

6. The most active lumber operations in the upper drainage basins of these rivers have been carried on within the last twenty years, although lumber began to be cut in large quantities as far back as the early seventies. Lumbermen state that in 1906 far more timber was cut in these drainage areas than in any other year on record. Before the seventies the forested areas in the mountain regions were practically the same as before the advent of the white settlers, deforestation having been slight up to that time; so that a comparison of the river conditions from about 1875 to the present day should show what changes, if any, have resulted from the reported wholesale destruction of our forests. In order to facilitate such comparisons I have had prepared a chart showing the rainfall records and river stages at three places on the Cumberland River, by months, and a similar one for the Tennessee River. These are transmitted herewith. The maximum and minimum gauge readings at each locality for each month have been plotted and joined by a line, to ascertain whether any noticeable change has taken place in the height or frequency of floods during the period under observation, or whether any marked difference in low waters or their frequency could be detected. There are also sent herewith charts showing the rainfall by years as far back as the records are available, to discover, if possible, whether any change is taking place in the quantity of precipitation, as it has been claimed by some that forests induce rainfall and that deforestation diminishes it.

7. I have examined these charts with minuteness, but can find no trace of any effect on the quantities of precipitation or on the fluctuations of stream flow that may be regarded as resulting beyond question from cutting off our forests. If any such effect has actually been brought about, it is so slight as to escape careful observation. The indications point in an opposite direction. 8. Efforts, thus far futile, have been made for years by scientific men to correlate the gauge heights of streams with local precipitation records by mathematical ratio, diagram, or curve, or in some other definite way, so that more or less exact predictions could be made therefrom. This has also been attempted by me from the data prepared for these rivers, but without any results that can be considered valuable. An examination of the tables will show some of the difficulties encountered. The experience of observers in the Weather Bureau also indicates that any such definite correlation would in all probability be unreliable and misleading. So many conditions enter into such a relationship that nothing satisfactory has yet been obtained in this direction.

9. It has been advanced, however, that some of our streams have deteriorated in navigable capacity, and that this has been in greater ratio than the diminution of rainfall in recent years, thus indicating, it is said, that forest protection is necessary. The arguments presented, when examined closely, often show evidences of "special pleading" and a neglect of pertinent facts. They are, in other words, too often open to the criticism that they have been brought forward for the purpose of "promoting a cause." The tables and diagrams necessary to support such efforts are usually based on wide general deductions, presumably drawn from the best precipitation records available, which, however, as is well known, are only kept at widely distant points and are frequently not characteristic or even representative. Their accuracy is also often questioned, as a strong wind may interfere with the exactness of the quantities shown by

the rainfall gauge, and at best they are far from being entirely satisfactory. The conclusions deduced from these scanty records are therefore by no means convincing. Nothing, in fact, can be more illusory than the theories sometimes derived from them, and nothing is more dangerous than to jump to wide generalizations from a few inadequate data. Some meteorologists think that the general variations in precipitation and other weather conditions occur in cycles, but the least period I have seen seriously suggested by them is about thirty-five years. Our records do not go far enough back to prove anything of value in this connection, and all efforts to establish run-off coefficients, curves, or diagrams, and formulas which might indicate the river stages that would result under given conditions of rainfall have thus far been practically useless; the lack of success in this direction arising mainly from our insufficient knowledge of all the conditions affecting the problem. Any positive or definite conclusions based on such uncertain data are necessarily very vulnerable.

10. It will be seen from the charts that the highest water at Nashville was that of 1882, and that high waters for the past ten years have been noticeably less in height and less in frequency than at the commencement of our records. At Chattanooga the highest water was in 1867. During the past few years, which have been rather "dry" years, the high waters in both the Cumberland and Tennessee have been less in height and duration than usual, which might reasonably be expected, owing to the smaller average quantities of precipitation; but it is particularly worthy of note that the low waters were not extreme, but as a rule were higher than usual, giving better depths for navigation. Although this may be in the nature of a surprise to some, it is plainly seen on the charts and is well established by reliable records. These facts can certainly be of no use in proving that our floods are higher and our low waters lower than before the forests were attacked. If they prove anything it should be the reverse. It may be stated broadly, and I think with positiveness, that the records of the Cumberland and Tennessee rivers do not indicate that the destruction of forests is having any noticeable effect, deleterious or otherwise, on stream flow or on precipitation.

11. The number of days and the number of times the river at Nashville has been above 35 feet or below 5 feet are shown in the accompanying table. In the last decade, 1899-1908, the stage has exceeded 35 feet but 8 times and for 78 days in all; for the previous decade it exceeded this height 9 times and for 123 days; and from 1879 to 1888, 11 times and 178 days. This plainly indicates a steady decrease of high-water stages in both duration and frequency since 1879. We also find that the river stage was less than 5 feet for 966 days in the years 1899-1904, an average of 161 days per year (in 1905-1908 low water was held up by Dam No. 1); in 1889-1898, 1.538 days, an average of 153.8 days per year: in 1879 -1888, 1,418 days, an average of 141.8 days per year, and in 1874-1878, 576 days, an average of 115.2 days per year.

By examining these averages we see that since 1878 the number of days the river was below 5 feet, or the duration of moderately low stages, has been increasing; but, on the other hand, in the years 1889-1904, the river reached a stage of less than 1 foot 3 times and for an average duration of 20.8 days per year; in the preceding decade the river was below 1 foot 7 times, with an average of 32 days per year; and in 1879 1888, 9 times, with an average of 32.5 days per year. It thus appears that extreme low-water stages have not been growing either in duration or frequency during the last thirty years. Any statements that low-water conditions are becoming worse on the Cumberland in more rapid ratio than the decrease in the rate of precipitation are untrue, for they are growing better.

12. The reason for some of this variation seems clear when we examine the record of rainfall. In the last decade the yearly average was 46.089 inches; in the next preceding, 50.969 inches; in the next, 56.326 inches; while from 1871 to 1878 it was 49,47 inches. This rule of change is thus seen to be substantially the same as for the duration of high-water stages, and may account in part for the low-water variations.

13. In the case of the Tennessee we find from the gauge records that the average number of days the river was above 25 feet was 7.0 from 1899 to 1908; from 1889 to 1898 the yearly average was 9.1; from 1879 to 1SSS it was 11.6; and for the four years before 1879 it was 9.3. These averages indicate that for thirty years the duration of stages over 25 feet has been steadily declining. Stages of over 35 feet occurred but four times during the ten years from 1899 to 1908, five times in 1889-1898, and 6 times in 1879-18SS. This shows that the

frequency of high water is also growing less. An inspection of the chart will shows that the heights are also less of late years, the highest flood having occurred in 1867 and the next highest in 1876. Moreover, we find that from 18991908 the river was below 1 foot only twice (1903, 1904), with an average of 10.3 days per year; from 1889-1898 only three times (1894, 1895, 1897), with an average of 7.7 days per year; whereas in 1879-1888 it was less than 1 foot five times, with an average of 20.4 days per year. These records demonstrate that low waters have neither been so frequent nor so low of late years. It is therefore not true that the high waters in the Tennessee River are growing higher and the low waters lower, nor is the frequency or duration of these stages increasing as claimed.

14. The frequency and height of high waters in both of these rivers, as well as the duration of low waters, follow the rule of precipitation closely enough during the periods under observation to force the conclusion that here is the principal cause of the variations in these streams, and that the introduction of forestry as an important factor in either direction is forced and illogical. Whenever extreme precipitation occurs the river immediately records its effect, and during seasons of light precipitation the corresponding result is to be expected. Temperatures play an important part as affecting evaporation and absorption by the watershed, rainfall of equal intensity at different times of the year not having the same effect on the river stages. The distribution of the rainfall, both in point of time and locality, also affects the river heights; but the extent of this feature is of course not apparent frmo the data available, owing to the small number of rainfall stations where records have been kept.

15. It is worthy of note, too, that there has been no falling off in late years in the navigable capacity of the Cumberland River or the Tennessee due to silting up of channels or any other cause. On the contrary, conditions affecting navigation are distinctly better now than twenty to thirty years ago, even in the portions unregulated. This is largely due, in all probability, to the removal of snags and surface obstructions and the clearing away of overhanging trees. The period of time covered by these observations and the extent of the decrease in forested areas seem ample to show any decided or important change that might be taking place. The records and facts, however, appear to indicate clearly that there is no such intimate or important relation between river stages and forest cover as have been so enthusiastically maintained by the advocates of this theory, nor has the soil wash of the upper regions been raising the beds. No one, I think, desires to detract from the value of forestry as forestry, but it has often been said that the random statements made as to the usefulness of forestry in improving the depths in our navigable streams have unfortunately been much overdrawn, and that the extravagant claims now being urged in this direction have not helped the development of this useful industry. Frequent repetitions may in time create a popular belief in almost anything, but can never alter facts.

16. The stages which interrupt navigation the most are the extremes of high and low. At either of these forest cover appears to be of little assistance, or perhaps even a detriment rather than a help. Forest cover, of course, delays somewhat the run-off of rains, and thus assists materially in promoting absorption by the ground. But these effects are only limited and of small value; the sponge is soon saturated by heavy rains, and floods then occur without hindrance. In dry weather the forests hold back the small rains, permitting them to evaporate more readily, and thus rob the streams of whatever help they might otherwise have obtained. As was recently pointed out by Col. H. M. Chittenden, Corps of Engineers (Proc. Am. Soc. C. E., Sept., 1908), it is fortunate that this important industry is not dependent on its relationship to navigable rivers for its usefulness. In their efforts to obtain some constitutional basis for their extensive plans, it is feared that the friends of this valuable department of the Government have imperiled their projects by seeking to furnish reasons for them which are not supported by the facts.

Very respectfully,

WM. W. HARTS, Major, Corps of Engineers.

INCREASE OF APPROPRIATION TO AGRICULTURAL COLLEGES.

HOUSE OF REPRESENTATIVES,

COMMITTEE ON AGRICULTURE,

Thursday, February 24, 1910.

The committee met at 11.45 a. m., Hon. Charles F. Scott in the chair.

There were present, in addition to the members of the committee, the following gentlemen interested in the pending bill: Dr. W. O. Thompson, president the Ohio State University; Dr. W. E. Stone, president Perdue University; Dr. J. L. Snyder, president Michigan Agricultural College; Dr. W. H. Jordan, director of the Geneva Experiment Station, New York; Dr. C. F. Curtis, dean of agriculture, Ames, Iowa; John Hamilton, farmers' institute specialist, United States Department of Agriculture.

The CHAIRMAN. The committee has met this morning pursuant to the order made yesterday to consider H. R. 15422.

[H. R. 15422, Sixty-first Congress, second session.]

A BILL For increase of appropriation to agricultural colleges for extension work.

Be it enacted by the Senate and House of Representatives of the United States of America in Congress assembled, That there shall be, and hereby is, annually appropriated, out of any money in the Treasury not otherwise appropriated, to be paid as hereinafter provided to each State and Territory, for the more complete endowment and maintenance of agricultural colleges now established or which may hereafter be established in accordance with the act of Congress approved July second, eighteen hundred and sixty-two, and the acts of Congress approved August thirtieth, eighteen hundred and ninety, and March fourth, nineteen hundred and seven, the sum of ten thousand dollars, in addition to the sum named in the said acts for the fiscal year ending June thirtieth, nineteen hundred and ten, and a like amount annually thereafter, to be applied by these colleges in giving instruction and demonstrations in agriculture, home economics, and similar lines of activity to persons not resident in these colleges in the several communities, as may be provided by the States accepting the provisions of this act, and in conveying and imparting to such persons information with reference to the improvement of rural life.

SEC. 2. That at any time after two years from the date on which any State or Territory has accepted the appropriation made by this act and has actually organized a separate and distinct department of extension work in connection with and as a part of its agricultural college there shall be available from the National Treasury, in addition to the ten thousand dollars herein appropriated for the purposes named in this act, an amount of money for each State and Territory equal to the amount appropriated by the State or Territory to its agricultural colleges for the current year for extension work: Provided, That the additional appropriation to any State or Territory shall not exceed an amount equal to one cent per capita of the total population of that State or Territory as shown by the last United States census.

SEC. 3. That all printed matter issued from the agricultural colleges for the furtherance of extension work, as provided in this act, shall be transmitted in the mails of the United States and dependencies free of charge for postage, under such regulations as the Postmaster-General may from time to time prescribe.

SEC. 4. That the sums hereby appropriated to the States and Territories for extension work shall be annually paid in equal quarterly payments on the first day of Janu-`