tion through the soil. Mr Berry's report shows that the season of 1899 was unusually windy, making evaporation greater than usual.

"Enough water was taken into canal No. 2 at Wheatland, Wyo., to have covered all the land irrigated to a depth of 2.53 feet, while only enough water was delivered through the J lateral of that canal to cover the two fields on which the water used in irrigation was measured to a depth respectively of 0.7 and 1.55 feet, the apparent loss in the canal being one-half the water entering it. In this case this high rate of loss is what might have been expected. The canal is long. It traverses a steep hillside slope for 2 miles, in which distance the loss under the lower bank is excessive. In many places the bottom is gravel, through which water escapes freely.

"In order to more carefully study the variations in these losses, arrangements were made early last season by Frank C. Kelsey, city engineer of Salt Lake City, Utah, to measure the seepage loss from the Jordan and Salt Lake Canal from the Jordan River. This canal is 29 miles long, with a bottom width of 20 feet. It originally had a grade of 2 feet per mile, but when measured was in bad condition, with a flow of 30 cubic feet per second at the head. The loss in 29 miles was 45 per cent. "The losses from seepage in new canals are excessive. For the past 6 months 500 inches of water have been flowing in at the head of a 10-mile lateral built at Billings, Mont., in 1899, but as yet not a drop has reached the lower end. On a canal built in Salt River Valley, Wyoming, there was a loss in 1896 of 10 cubic feet per second in a distance of 100 feet, which continued for several weeks with no apparent prospect of the loss diminishing. This was about one-third of the canal's flow. The canal was then abandoned. The canals which take water from the North Platte River are all subject to excessive losses when first built, because of the sandy soil through which they must pass. In high water, however, this river is heavily charged with a white clay, due to the erosion of its banks. When this is deposited on the sides and bottom of ditches it forms a coating only less impervious than cement, and after a few weeks' operation during high water seepage losses always show great diminution.

"Mr. Code reports that the water of Salt River, Arizona, contains a cementing material which in time renders its banks almost water-tight so long as they remain undisturbed. This has not heretofore been possible on the Mesa Canal, because it has been undergoing constant repairs and improvements."

Q. (By Mr. A. L. HARRIS.) Does that seepage come to the surface below on ground that may be used for crops?-A. Yes. The loss of water by seepage is not only a serious problem with canal owners, but frequently becomes the cause of grave injury to the farming lands below. The water which escapes through the bottom of the canal follows the path of least resistance, and this sometimes takes it into the channel of the river or causes the appearance of springs in ravines which before were dry, or it may lead it to reappear in the fields below, often converting them into marshes and swamps. Instances are not infrequent where thousands of acres of land have for a time been rendered valueless from this cause. The saturation of the subsoil and the gradual rise of the water level nearly always attend irrigation. The first wells dug in the San Joaquin Valley in California were 60 to 70 feet deep. Since then the water has risen in many of these wells to within 4 or 5 feet of the surface. Where seepage is not excessive it furnishes an inexpensive method of irrigation; where it is it may cause a double injury. It prevents the growth of crops because of too much water and rendors the soil unproductive through the accumulation of alkali which it causes. Water passing from canals through the subsoil dissolves the soluble salts which all Western lands contain in greater or less measure, and the subsequent evaporation of this alkali-impregnated water so increases the percentage of alkali in the lower lands as to prevent the growth of crops. This evil is not, however, destined to be a permanent one, and, like the excessive moisture, can be remedied by drainage.

Q. (By Mr. FARQUHAR.) In Colorado and Wyoming is the general characteristic of your streams seepage or are they on solid ground and solid bottom?-A. In both of those two States, as a rule, the losses are not excessive in canals. There are exceptions, of course, but in both eastern Colorado and eastern Wyoming the soil is of a character to hold water pretty well, although in the older districts in Colorado there is a considerable area of land in the low bottoms along the streams that has become supersaturated both with water and alkali. This is not altogether due to losses from ditches. Probably the greater part comes from putting too much water on the fields. It is rather a drainage from the area irrigated.

Q. The characteristics of the Arkansas from its source north of Leadville to its reaching the Mississippi has usually been characterized as a river of seepage. Any. where along the banks by digging a few feet down you reach a well. Is it not a fact that that river itself, in its whole course, a good part of it, is really below the surface?-A. All of the rivers that flow out on the plains sink into the sand of their bed. I did not take your question as applying to the rivers, but rather to the ditches. Q. The point was this, that many of the foothill streams east of the mountains

would not the seepage be generally supposed to amount to a great deal or do much harm, but when you come to a river like the Arkansas, with a large body of water passing over plains with very little fall, the water itself is drawn out and distributed a good deal in the banks and surrounding low ground?-A. That is true of all the streams flowing from the Rocky Mountains out on the plains. It is a characteristic of the Rio Grande, of the Arkansas, the South Platte, and the North Platte. The North Platte has been measured 100 miles west of the Wyoming border and found to carry 400 cubic feet per second, while a few miles east of the Nebraska and Wyoming border it was entirely dry. The entire 400 cubic feet per second had sunk into the sand.

Q. Have you anything to say about the filling up of these canals with silt and other substances that are quite expensive in canaling?—A. The canals taken out of the lower portion of those streams running out on the plains are more or less troubled by the moving sands in the bottom of the stream that tend to fill them up; and all canals that are taken out of rivers that carry considerable quantities of mid in high waters have to be cleaned out every year. The deposits of mud can be handled as a rule without any excessive expense, but in streams like the North and South Platte and the lower part of the Arkansas the sand question is quite troublesome; and on the lower part of the Rio Grande the question of mud becomes an important factor. The red rise in the Rio Grande occurs when there are torrential rains along certain portions of the river where there is a red soil, and enormous volumes of mud are washed down in the river; samples of the stream taken at that time have shown as high as 17 per cent of solid matter. All the ditches have to be closed during the time of the red rise because they would immediately fill up.

At Las Cruces, N. Mex., is one of the oldest ditches, if not the oldest ditch, in the United States. That ditch was formerly a channel cut below the surface of the ground. Now it is raised 4 or 5 feet above the surface of the ground. As the mud which was carried into the ditch was cleaned out each year it was thrown on the banks; and when the banks became so high as to be troublesome they simply let the mud fill up a foot or so in the bottom. In time the ditch got above the stream, and they had to move the head farther upstream. In the period of operation of that canal the head has been moved upstream 3 or 4 miles from the original location. Not only that, but since the time the irrigation began the level of the soil on which the water and mud has been spread has been raised from a few inches to 2 feethigher, of course, nearer the ditch, and becoming thinner and thinner as you recede from it. The Rio Grande at El Paso has filled up its channel from this cause until the river itself is higher than some of the streets of either El Paso or Juarez.

Q. (By Mr. A. L. HARRIS.) Has there been any conflict between irrigators on account of priority of rights?-A. In recent years litigation and controversy over the division of water has been alike a conspicuous and injurious feature of our irrigated agriculture. It has been due to two causes. The first is the lack of any plan for the establishment of rights to a stream, or public protection of those rights. When the men along the lower end of a stream see its waters shrink and their crops burning up for the lack of water they realize that it is due, not to the absence of the snow in the mountains, but to the fact that later ditches above them are robbing them of their just share. Before the farmers will permit the loss of their year's labor from this cause they will resort to almost any expedient to obtain what they believe belongs to them, and so they organize raids to tear out the dams above or go into court to obtain legal redress. The remedy for this is to have water divided under public control. In the 4 States where this has been done irrigators are far more contented and controversies far less numerous and injurious than where no such control has been exercised. The second reason for controversies has grown out of the mistakes made in the adjudication of rights to streams. In the study of the water-right problems of California recently completed there were claims for 28,630,932 inches from a stream which can not be relied upon to furnish 10,000 inches. On another stream which carries in the irrigation season less than 200 cubic feet of water per second there were claims amounting in the aggregate to 147,600 cubic feet per second. On another river whose greatest measured flow is less than 60,000 cubic feet per second there are claims amounting in the aggregate to 461,794 cubic feet per second in addition to six separate claims of the entire supply.

The situation in California is the situation in nearly every other arid State or Territory. Before the value of water was appreciated titles to its use or control in amounts far beyond the present or any possible future need of appropriators were repeatedly established, and the question whether these excess rights are now to be corrected or to be recognized as vested rights is one of the grave issues confronting irrigators, lawmakers, and courts in every Western State.

In 1884 and 1885, while acting as assistant State engineer of Colorado, I measured the ditches of northern Colorado on the streams which had been previously adjudicated. My report of these measurements called attention to the discrepancy between the decreed appropriations and the actual carrying capacity of these ditches and canals in the following terms:

"So great was this in some instances that the result of the gaugings and the decreed capacity seemed to have no connection with each other. Ditches were met with having decreed capacities of two, three, and even five times the volume they were capable of carrying, ever have carried, or will probably ever need. Other ditches in the same district have decrees which fairly represent their actual needs. It needs no argument to show the worse than uselessness of these decrees as a guide to the water commissioner in the performance of his duties."

When these decrees were rendered the majority of appropriators believed that rights for irrigation were limited to the lands already irrigated, and that, so long as used there, the actual volume stated in the decree cut very little figure. Hence there was little solicitude on the part of late appropriators as to any danger arising out of excessive grants. Under the terms of these decrees each appropriator is entitled to a definite volume of water, described in cubic feet per second, and to a concinuous flow of this volume throughout the year.

Recent decisions have recognized the right of the holders of these decreed appropriations to sell the entire volume granted. As a result, the owners of earlier pri.

Relation between the mean monthly discharge of the Poudre River and the appropriations therefrom. orities are enlarging their ditches and extending them to other lands, or, where this is not possible, are attempting to dispose of the surplus to other users. Every attempt to do this, however, is contested. The truth is that irrigators have, in practice, been building up a system of one theory of water rights while the courts have rendered a number of decisions based on another theory. We have now reached a point where one of the two must give way. If the doctrine laid down in these decisions is carried to its logical conclusion it will transfer the ownership of a majority of the

streams of northern Colorado to a few early appropriators, and compel a large preportion of the actual users of water to purchase from such appropriators the wate they have heretofore had for nothing. That this is not an extreme statement » shown by the accompanying diagram, which exhibits the relation between the mean monthly discharge and the decreed appropriations of the Poudre River.

The last examination of the records showed there were 104 appropriators from this river, the aggregate of these rights being 4,632 second-feet, each right being for a continuous discharge of the volume decreed; yet in Angast of 1834 the stream carried only 162 cubic feet per second; in August, 1833, 141 second-feet, and the stream has frequently fallen during the irrigation season to below 100 second feet. If the holders of these rights had lived up to their opportunities during the last half of their irrigation season, fully one-half of the actual users of water would have had to buy from the holders of these excess rights every gallon of water used after the middle of August. That they have not been compelled to do this is due to the fact that irrigation practice in that State is superior to irrigation law.

The appreciation of the dangers which this situation creates is not confined to farmers alone. In a different brief from the one before referred to it is thus forcibly stated by Judge Elliott:

"Excess priority decrees are a crying evil in this State. From every quarter the demand for their correction is strong and loud. Such crying demand can not be silenced by declaring that the meaning and effect of such decrees can never be inquired into, construed, or corrected after 4 years.

"In many cases such decrees are so uncertain, so ambiguous, so inequitable, so unjust, and their continuance is such a hardship that litigated cases will be continnally pressed upon the attention of the courts until such controversies are heard and settled, and settled right. Litigation in a free country can never end while wrongs are unrighted."

The settlement of this issue is not of local importance. It concerns the State and nation as much as the individual irrigator. The individual irrigator needs to know who owns the water he uses, if State or national aid is to be extended. It needs to be known who owns the water which public funds render available.

Before either public or private development proceeds much farther there is need of some more general agreement regarding the nature of a water right than now prevails, as well as some more effective means of disposing of streams than has yet been provided. For several years Canada has been dealing with this problem and has finally reached a definite result. The fact that their conditions are similar to ours makes the general principles which underlie the Canadian irrigation code worthy of our study. These principles are given below:

"(1) That the water in all streams, lakes, ponds, springs, or other sources is the property of the Crown.

(2) That this water may be obtained by companies or individuals for certain described uses upon compliance with the provisions of the law.

"(3) That the uses for which water may le so acquired are 'domestic,' 'irriga. tion,' and 'other' purposes, domestic purposes being limited to household and sanitary purposes, the watering of stock, and operation of railways and factories by steam, but not the sale or barter of water for such purposes.

"(4) That the company or individual acquiring water for irrigation or other purposes shall be given a clear and indisputable title to such water.

"(5) That holders of water rights shall have the protection and assistance of permanent Government officials in the exercise of such rights.

"(6) That disputes or complaints regarding the diversion or use of water shall be referred to and settled by the officials of the Government department charged with the administration of the act, and that decisions so given shall be final and without appeal."

It is interesting to compare these principles of the Canadian law with those underlying the Wyoming irrigation code, Wyoming having gone farther than any of the other arid Commonwealths in direction of public control of streams. These follow: "First. That water is not subject to private ownership, but is the property of the State.

"Second. That the board of control is the trustee for the administering of a great public trust in the interests of the people of the State.

"Third. That all rights to divert water from the streams must be based on beneficial use, and that the right terminates when the use ceases.

"Fourth. That the volume diverted shall in all cases be limited to the least amount actually necessary for the accomplishment of the purposes of the diversion.

"Fifth. That under no circumstances shall the water diverted for irrigation exceed

1 cubic foot per second for each 70 acres of land actually irrigated.

1 Bul. 58, Office of Experiment Stations, U. S. Department of Agriculture, pp. 30-32.

Bul. 96, Office of Experiment Stations, U. S. Department of Agriculture, p. 12.

"Sixth. That the right to the use of the public waters attaches only to the use for which the right was originally obtained.

"Seventh. That the right of diversion for irrigation attaches to the land reclaimed and none other; that the transfer of the land carries with it the right, and that apart from the land the right can not be transferred.

"Eighth. That when a ditch waters land not the property of the ditch owner the right attaches to the land on which the water is used and not to the ditch. The owner of the lands irrigated makes the proof of appropriation and the certificate is issued to him. No certificate of appropriation can be issued to a ditch owner for the watering of lands not his own. The ditch owner is a common carrier and is subject to regulation as such.

Ninth. That when proper diligence has been exercised in the construction of works and in applying the water to the purpose for which it is diverted the priority is fixed by the date of beginning the survey. When diligence is lacking, the priority dates from the time of use."

Q. (By Mr. FARQUHAR.) Does the State control the reservoirs?-A. Except for 2 or 3 reservoirs in Colorado, all the reservoirs there are in the West are private property, and their owners exercise the same control over them that they do over ditches. Irrigation from reservoirs has not yet, however, assumed much importance as compared to irrigation from canals which take water directly from the streams. So long as there is water running in a river which can be diverted there is no need of reservoirs, since storage is only an added expense to the direct diversion from streams. On every river, therefore, reservoirs receive little attention until the natural flow has been utilized; that is, on rivers having a perennial flow. On the Pondre River in Colorado, however, the natural flow has been exhausted and an extensive system of private reservoirs have been built to supplement it.

Q. (By Mr. KENNEDY.) How far is Greeley from the head of the stream that feeds their canal?-A. About 125 miles from the head; about 40 miles below the head of the upper ditch.

Q. Are there lands adjacent to that river all the way to the head, as well as the Greeley Colony lands?-A. Only to where the stream leaves the mountains, about 40 miles above.

Q. Plenty of water for all?-A. No; there is plenty of water early in the season, but they have had to resort to storage to secure enough for the later part of the

season.

Q. (By Mr. A. L. HARRIS.) Is there a possibility for storage so as to economize the water at dry seasons?-A. Yes.

Q. What is the effect, if any, of storage on prior rights?-A. That is a troublesome question to answer.

We will take up the question of storage in connection with this matter. Bulletin 92 deals with the subject of storage on the Poudre River. It is the stream where storage has been carried further, probably, than any other in the whole Rocky Mountain drainage area. The diagram of the run-off of that river in the different months of the irrigation period for a large number of years shows that distribution of water during the season is far from uniform. The highest water occurs in May, and from the middle of May to the middle of June nearly half of the entire year's discharge runs away. The needs of irrigators in this valley are not in accord with this variation in discharge. They are now growing crops which require more water in July, August, and September than the stream will supply, and this has made it necessary for them to build reservoirs to hold back the surplus flow of the early summer months until it is needed. They have done this by utilizing natural depressions which lie outside of the channel of the stream, which are filled by the higher canals and turned into the lower ones. The development of this reservoir system has given rise to a very interesting system of exchanges between the canals, described in this bulletin, and hence need not be referred to here. Where reservoirs are located outside of the channel of streams there is no question of public policy involved in their construction and operation as private works, and as irrigation extends there will be more and more private capital invested in such reservoirs, because these investments are proving exceedingly profitable. But where reservoirs are located in the channel of running streams, and especially in the mountains on the headwaters of these streams, there is a question of public policy as to whether or not they should be built as public works, even if private capital is willing to undertake their construction. The water from reservoirs so located has to be turned into the natural channel of the stream and carried down with the natural flow to the valleys where it is to be used. If there is no public control of streams, irrigators will not discriminate between the natural flow and the stored supply. They will raise their head gates and take whatever comes in. Unless there is some means for the public regulation of head gates, those having no right to the stored supply will often times have a better opportunity for

1 Bul. 96, Office of Experiment Stations, U. S. Department of Agriculture, pp. 49 and 50.